Unplanned 30-day hospital readmission is an economic burden on the healthcare system, with an estimate cost to Medicare around $17.4 billion a year. Patients readmitted after colectomy or proctectomy are subject to a high mortality rate and increased complications1,2. In an effort to reduce readmission rates, the Center for Medicare and Medicaid Services (CMS) instituted a new policy in 2012, which established the Hospital Readmission Reduction Program. This program required CMS to reduce payment to hospitals that have a higher readmission rate than the national average, thus establishing readmission rate as a key factor in performing high-value, cost-conscious care3. Thirty-day readmission and mortality rates for specific diseases are now considered public knowledge in order to increase quality of care performed during inpatient admissions and to assist patients in selecting a hospital in which they would prefer to receive care4.
Ulcerative colitis (UC) is a chronic, relapsing disease with a prevalence of 238 per 100,000 patients per year in the United States5. Due to medically refractory disease, chronic corticosteroid use, untoward side effects of medication, dysplasia, or colorectal cancer, ˜5–20% of UC patients undergo colectomy or proctectomy within 10 years of being diagnosed5–7. After undergoing colectomy or proctectomy, patients are at an increased risk for post-operative complications with 17–32% of patients having an unplanned readmission within 30 days8–10. Previous studies have evaluated factors associated with readmission, and while some have not determined any significant predictors, other studies have found proctocolectomy with ileal pouch-anal anastomosis (IPAA) and diverting ileostomy, age, and post-operative complications to be associated with readmission. However, these studies were all conducted on small patient populations at single institutions8,10,11.
Thus, we sought to determine the 30-day readmission rate and associated predictors of readmission in UC patients undergoing colectomy or proctectomy utilizing a large national clinical database. We hypothesized that many risk factors including pre-operative conditions, type of surgery, and post-operative complications would impact 30-day readmission rates.
Materials and methods
Data were obtained from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) patient user file between 2011 and 2015. The ACS NSQIP collects reliable and validated information by trained and audited surgical reviewers at ˜700 hospitals across the country. Data from the pre-operative period through 30-day post operation is collected using an 8th-day sample methodology in order to ensure that cases have an equal chance of being selected. Information obtained for each patient included patient demographics, laboratory values, medications, comorbidities, and 30-day post-operative outcomes and complications for patients undergoing a variety of surgeries12,13. Missing information from the user file during the study period is outlined in Supplementary Table 1.
Study population and definition of variables
The study population included adult patients with a diagnosis of UC that underwent colectomy or proctectomy between 2011 and 2015. These patients were selected with the use of International Classification of Disease (ICD-9 and 10) codes for UC (556.xx and K51.xx, respectively) and Current Procedural Terminology (CPT) codes for colectomy or proctectomy. We excluded patients who died during the hospitalization or were hospitalized for >30 days.
Definition of pre-operative comorbidities, medications, and laboratory values
Pre-operative comorbidities included obesity, diabetes mellitus, severe chronic obstructive pulmonary disease (COPD), ascites, congestive heart failure within 30 days before surgery, hypertension requiring medication, acute renal failure, corticosteroid use for chronic conditions, >10% loss of body weight in the 6 months before surgery, bleeding disorders, anemia requiring transfusion prior to surgery, tobacco abuse within the year before surgery, and American Society of Anesthesiology (ASA) classification.
Pre-operative laboratory values included aspartate transaminase, hematocrit, albumin, and alkaline phosphatase.
Definition of colectomy or proctectomy and other surgical characteristics
Colectomy or proctectomy was defined as one of the following procedures: abdominal colectomy with end ileostomy (CPT codes: 44150 and 44210), proctocolectomy and end ileostomy (CPT codes: 44155 and 44212), proctocolectomy with IPAA creation and diverting ileostomy (CPT codes: 44157, 44158, and 44211), and proctectomy, which included proctectomy with combined abdominal perineal pull-through with creation of colonic reservoir and diverting enterostomy (CPT codes: 45119 and 45397), and proctectomy, partial with rectal mucosectomy, ileoanal anastomosis, and creation of ileal reservoir, with or without loop ileostomy (CPT code: 45113). Although proctectomy with combined abdominal perineal pull-through with creation of colonic reservoir and diverting enterostomy (CPT codes: 45119 and 45397) is not a common surgery for UC, many hospital coders including our institution coders often used it to code proctectomy with IPAA creation. In order to maximally capture patients, we included CPT codes, 45119 and 45397. Other surgical characteristics included laparoscopic versus open surgery, operation time, emergent versus non-emergent surgery, length of stay (LOS) associated with index surgery, and return to the operating room during the same admission.
Definition of post-surgical complications
Post-operative complications included surgical site infection (defined as superficial surgical site infection, deep incisional surgical site infection, and organ space surgical site infection), wound disruption, pneumonia, pulmonary embolism, renal insufficiency, urinary tract infection (UTI), cerebral vascular accident with neurologic deficits, cardiac arrest requiring cardiopulmonary resuscitation or myocardial infarction, bleeding transfusions, deep vein thrombosis (DVT) or thrombophlebitis, and sepsis or septic shock.
SAS 9.4 (SAS Institute, Cary, NC) was used to perform all analyses. Continuous data were summarized with means and standard deviations and differences between groups were evaluated with t tests; log-transformations were employed where necessary prior to performing the t test. ×2-tests and Fisher exact tests were used to determine differences in categorical, as appropriate. Using NSQIP data from 2011 through 2014 as a developmental dataset logistic regression model was fit for 30-day readmission to evaluate factors association with readmission. Terms included in the model were determined through stepwise selection. This logistic regression model, with all continuous variables dichotomized, was used to construct the risk score for readmission. The coefficients, multiplied by 10 and rounded to the nearest integer, provided the weights used in the score for each of the risk factors. The fit of the model developed with NSQIP data from 2011 to 2014 was externally validated with NSQIP 2015 data. Statistical significance was defined by p value <0.05.
Thirty-day readmission rate
A total of 4797 UC patients that underwent colectomy or proctectomy were included in the study. The 30-day readmission rate was 20% (N = 963).
Table 1 is the univariate analysis of baseline patient characteristics, Table 2 is the univariate analysis of baseline patient laboratory results, and Table 3 is the univariate analysis of operative characteristics and complications. These tables summarize the differences in baseline characteristics between those who were and were not readmitted. Hispanic patients, patients with pre-existing congestive heart failure, hypertension requiring medication, or elevated alkaline phosphatase, patients undergoing proctocolectomy with IPAA creation and diverting ileostomy, patients with an unplanned return to the operative room during index admission, longer operative time, or certain postoperative complications had significantly higher rates of 30-day readmission. Table 4, which is the subgroup analyses by surgery performed of emergency admission and operative time, summarizes the results of evaluating emergency cases and operation time. Emergent surgery did not impact 30-day readmission rates on patients receiving abdominal colectomy with end ileostomy.
Multivariate analysis for factors associated with readmission
Table 5, which is the multivariate logistic regression model for readmission within 30 days, summarizes risk factors associated with 30-day readmission. Patient's ethnicity, such as Hispanic race (odds ratio (OR) 1.66, 95% confidence interval (CI) (1.11, 2.50)), and certain preexisting conditions before surgery, such as congestive heart failure (OR 4.76, 95 % CI (1.09, 20.75)), lower albumin (OR 0.81, 95% CI (0.70, 0.93)), hypertension requiring medications (OR 1.43, 95% CI (1.12, 1.82)), and utilization of corticosteroids (OR 1.27, 95% CI (1.02, 1.57)) significantly impacted 30-day readmission rate in UC patients undergoing colectomy and proctectomy.
Type of surgery and other factors associated with surgery also significantly impacted 30-day readmission rates. Patients were more likely to be readmitted if they underwent proctocolectomy with IPAA creation (OR 1.53, 95% CI (1.10, 2.10)) or other surgeries including proctectomy with combined abdominal perineal pull-through with creation of colonic reservoir and diverting enterostomy, and proctectomy, partial with rectal mucosectomy, ileoanal anastomosis, and creation of ileal reservoir, with or without loop ileostomy (OR 1.69, 95% CI (1.18, 2.42)). Return to the operation room (OR 3.71, 95% CI (2.56, 5.40)), longer operation time (OR 1.04, 95% CI (1.01, 1.07)), and shorter LOS (OR 0.93, 95% CI (0.91, 0.95)) were associated with an increased risk of 30-day readmission. Post-operative complications including surgical site infection (OR 3.38, 95% CI (2.59, 4.42)), pulmonary embolism (OR 4.34, 95% CI (1.72, 10.91)), renal insufficiency (OR 3.38, 95% CI (1.51 7.54)), UTI (OR 2.83, 95% CI (1.86, 4.31)), DVT/thrombophlebitis (OR 5.69, 95% CI (3.67, 8.83)), and sepsis/septic shock (OR 2.22, 95% CI (1.57, 3.12)) were associated with higher 30-day readmission rates.
The ability of the logistic regression model to discriminate between patients who will be readmitted within 30 days and those who will not was evaluated with the area under the receiver operating characteristic curve (AUC). The AUC for the model fit with the developmental dataset was 0.73 (95% CI (0.71, 0.76)) demonstrating good discrimination. When the derived model is applied to the validation dataset, the discrimination of the model remains good with a similar AUC of 0.71 (95% CI (0.66, 0.75)).
Risk score to predict readmission
Table 6 outlines a risk score formulated in order to determine an individual's risk for readmission. The range of risk score in the study was 0–58. Figure 1 describes proportion of patients readmitted. Ten percent of patients with a risk score between 0 and 9 were readmitted, 18.5% with a score between 10 and 19, 52.2% with a score between 20 and 29, and 59.6% in patients with a risk score >29. The average risk score for patients included in the study was 11.8 with a standard deviation of 8.53. In Table 7, the model is validated with NSQIO 2015 data. When the model is applied to the validation dataset, a similar trend was observed. The average risk score for patients was 11.7 with a standard deviation of 8.15.
UC is a chronic relapsing disease. Once patients have medically refractory UC, untoward side effects from medications, or develop dysplasia or malignancy, surgery is indicated. In this study, utilizing a large national clinical database, 20% of UC patients undergoing colectomy or proctectomy were readmitted within 30 days. While multiple factors were associated with readmission, factors with the highest association were potentially modifiable, including post-operative complications and uncontrolled pre-existing conditions. This study also established a risk score model to predict readmission in UC patients that undergo colectomy or proctectomy; therefore, by applying to each UC patient undergoing colectomy or proctectomy, the model will be able to determine patients with high risk of readmission. By targeting modifiable risk factors, healthcare providers might be able to reduce the 30-day readmission.
In our study, multiple pre-existing conditions of were associated with readmission. Hispanic patients with UC that undergo colectomy or proctectomy were more likely to be readmitted within 30 days. Hispanic patients with UC undergoing surgery have a higher risk of post-operative complications, and are more likely to be readmitted after surgery14. However, our study controlled for many post-operative complications, suggesting that other factors also contribute to the high 30-day readmission rate. Hispanic patients with UC more likely have increased disease severity and decreased use of biologics, which might partially contribute to the increased readmission after colectomy or proctectomy15.
UC patients with lower pre-surgical serum albumin, indicating malnutrition, were more likely to be readmitted. The impact of malnutrition on surgical outcomes has been well studied and is associated with increased complications, mortality, and readmission in abdominal surgery16; therefore, if surgery is not emergent, nutrition should be optimized prior to surgery to reduce 30-day readmission rates in UC patients.
Pre-existing heart failure and hypertension requiring medication were also associated with an increased risk of readmission in UC patients undergoing colectomy or proctectomy. Patients with heart failure that undergo non-cardiac surgery are at a high risk of complications, with 30% of patients experiencing myocardial infarction, exacerbation of pre-existing heart failure, or other complications that would prompt readmission17. In addition, patients with hypertension requiring medications were more likely have an unplanned readmission after receiving plastic surgery18. UC patients with heart failure and hypertension should be optimized before undergoing surgery to reduce the risks of readmission.
Chronic use of corticosteroids in patients with UC is associated with an increased risk of readmission, which may be related to an increased risk of post-operative complications and infections19. However, our study controlled for these variables, suggesting that other factors including increased disease severity in patients with chronic corticosteroid use may contribute to this as well. Every effort should be made to minimize corticosteroid use prior to surgery in UC patients to reduce the risk of 30-day readmission.
While multiple surgical techniques can be utilized to treat UC patients, these procedures can be broadly categorized into restorative and non-restorative procedures. Restorative procedures involve creation of a small intestinal pouch with restoration of natural bowel function, whereas non-restorative procedures involve creation of a permanent colostomy. Since restorative procedures involve creation of small intestinal reservoir and a low pelvic anastomosis, these procedures entail a higher degree of technical complexity and longer operative times compared to non-restorative procedures. Additionally, restorative procedures are associated with increased complications when compared to other types of colectomy20. As expected, our data indicated that restorative surgeries were associated with increased 30-day readmissions21. Though patients that undergo an abdominal colectomy with ileostomy can have complications, such as delayed wound healing2, patients that undergo proctocolectomy with IPAA creation have approximately a 20% chance of bowel obstruction and 5% chance of pelvic sepsis, which leads to an increased risk of readmission20. Given the high risk of complications associated with restorative procedures, many centers are now performing these surgeries in three stages to allow cessation of immunosuppression before the creation of intestinal pouch. Furthermore, given the technical complexity and complicated post op care of these patients, these procedures are best performed at high volume, specialized centers to minimize post-operative complications and resultant readmission22.
UC patients that experienced an unplanned return to the operating room during their index admission were more likely to be readmitted within 30 days and this trend has also been noted in patients that have undergone other surgeries, for instance, lower extremity arterial bypass surgery23. Longer operation time was associated with increased risk of readmission in UC patients undergoing colectomy or proctectomy in our study, which is in agreement with a prior study in patients that undergo plastic surgery24. Previous studies have noted an exponentially increased risk of complications as surgery length of time increased by 1–3 h intervals24.
UC patients, especially patients who have multiple comorbidities or fail to respond to medical therapy, are at an increased risk of post-operative complications after colectomy or proctectomy9. Post-operative complications were associated with increased risk of readmission not only in UC patients undergoing colectomy or proctectomy, but also in other patients undergoing other surgeries25. Given the increased financial burden of postoperative complications and 30-day readmission, healthcare providers should aim to reduce short-term complications to improve patient and hospital outcomes, including minimizing intraoperative fecal contamination, optimal DVT prophylaxis, and early removal Foley catheters following surgery26,27.
Shorter LOS was also associated with an increased risk for 30-day readmission in patients with UC that undergo colectomy or proctectomy. Similarly, patients with other chronic diseases, including heart failure and COPD with decreased hospital LOS have increased 30-day hospital readmission rates28,29. Previous studies on patients receiving care at the Veterans Association (VA) hospital projected that for each day, hospital LOS was reduced, there was a 6% chance of readmission28. The association of increased hospital 30-day readmission rate with decreasing index hospital LOS raises concerns about quality of care that may be affected by the interventions that have been instituted to reduce hospital costs and LOS30. While minimizing the economic burden of healthcare in society is crucial, focus must also be placed on the quality of care and safe discharge planning, with initiative such as the Enhanced Recover After Surgery, which has successfully decreased LOS, mortality, and readmission rates31.
This study has multiple strengths. Recognizing risk factors that are associated with readmission are crucial in order to determine interventions that can reduce this vulnerable patient population from unnecessary readmissions. We established the first risk score model in order to identify UC patients who are at the highest risk of readmission after colectomy or proctectomy. Utilizing this model will allow physicians to target and intervene on patients. This will not only improve hospital outcomes but patient outcomes. In addition, this knowledge will help tremendously in setting patient expectations and perioperative planning. Utilization of the ACS NSQIP database allows the collection of data on a large number of UC patients who underwent colectomy or proctectomy, which could not be completed from a single medical center study. Finally, this database includes a diverse patient population from different geographic regions of the United States allowing for an accurate representation of the entire country.
However, this study has several limitations. Due to the cross-sectional analysis utilized in this study, only association between risk factors and readmission rates could be determined. However, in order to determine causality, a prospective analysis must be performed. In addition, this study was unable to determine whether the patient was taking other medications, including immunomodulators and biologic medicines prior to surgery, which could also impact 30-day readmission. These medications are associated with immunosuppression, renal dysfunction, and other complications, which would make patients taking these medications more prone to develop specific complications, including infections and acute kidney injury; therefore these patients would have a higher risk of readmission32. This study also does not evaluate initiatives such as enhanced recovery, infection prevention bundles, extended venous thromboembolism prophylaxis or readmission reduction pathways, which have been shown to reduce readmission in other patients31.
This study utilized a large national clinical database to determine factors that are associated with 30-day readmission in patients with UC that undergo colectomy or proctectomy. Pre-existing conditions, corticosteroid use, and post-surgical complications were associated with increased 30-day readmission rates; therefore, optimization of the patients’ chronic medical conditions, nutritional status, medication, and prevention of post-surgical complications may decrease rates of readmission. Patients with UC undergoing colectomy or proctectomy might benefit from receiving staged surgeries at a tertiary care center with skilled surgeons to reduce readmissions. Focus on quality of care while decreasing LOS should be emphasized in order to decrease the 30-day readmission rates. Many factors associated with readmission are potentially modifiable and these should be targeted to improve readmission rates in this vulnerable patient population. Overall, 30-day readmissions have a major effect on patients, providers, and healthcare costs, and further research should be conducted to determine ways to reduce readmission rates.
WHAT IS CURRENT KNOWLEDGE
- Thirty-day hospital readmission in patients undergoing surgery has a significant economic burden on the healthcare system.
- Approximately 5–20% of patients with ulcerative colitis (UC) undergo colectomy or proctectomy within 10 years of diagnosis and are at an increased risk of complications prompting unplanned hospital readmission.
WHAT IS NEW HERE
- The 30-day readmission rate was 20% in UC patients receiving colectomy or proctectomy.
- There are multiple potentially modifiable risk factors that are associated with high 30-day readmission rate.
- Intervention aimed at these risk factors may improve outcomes of these patients.
We greatly appreciate the guidance and suggestions from the Mentored short clinical research project program of the Center for Faculty Advancement, Mentoring and Engagement (FAME) at the Ohio State University.
1. Jencks, S. F., Williams, M. V. & Coleman, E. A. Rehospitalizations among patients in the Medicare fee-for-service program. N. Engl. J. Med.360
, 14180-21428 (2009).
2. Kazaure, H. S., Roman, S. A. & Sosa, J. A. Association of post discharge complications with reoperation and mortality in general surgery. Arch. Surg.147
, 1000-1007 (2012).
5. Kappelman, M. D. et al. The prevalence and geographic distribution of Crohn's disease and ulcerative colitis in the United States. Clin. Gastroenterol. Hepatol.5
, 1424-1429 (2007).
6. Frolkis, A. D. et al. Risk of surgery for the inflammatory bowel diseases has decreased over time: a systematic review and meta-analysis of populationbased studies. Gastroenterology145
, 996-1006 (2013).
7. Cima, R. R. Timing and indications for colectomy in chronic ulcerative colitis: surgical consideration. Dig. Dis.28
, 501-507 (2010).
8. Feuerstein, J. D. et al. Surgery for ulcerative colitis is associated with a high rate of readmissions at 30 days. Inflamm. Bowel Dis.21
, 2130-2136 (2015).
9. Silva, S. D. et al. Postoperative complications and mortality following colectomy for ulcerative colitis. Clin. Gastroenterol. Hepatol.9
, 972-980 (2011).
10. Gu, J. et al. Factors associated with postoperative morbidity, reoperation and readmission rates after laparoscopic total abdominal colectomy for ulcerative colitis. Colorectal Dis
11. Frolkis, A. et al. Postoperative complications and emergent readmission in children and adults with inflammatory bowel disease who undergo intestinal resection. Inflamm. Bowel Dis.20
, 1316-1323 (2014).
12. Hall, B. L. et al. Does surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals. Ann. Surg.250
, 363-376 (2009).
13. Ingraham, A. M. et al. Quality improvement in surgery: the American College of Surgeons National Surgical Quality Improvement Program approach. Adv. Surg.44
, 251-267 (2010).
14. Carthon, J. M. B. et al. Variations in postoperative complications across race, ethnicity and sex among older adults. J. Am. Geriatr. Soc.61
, 1499-1507 (2013).
15. Lattimer, L. D., Chandler, M. B. & Borum, M. L. Hispanics and inflammatory bowel disease. Inflamm. Bowel Dis.21
, 1214-1218 (2015).
16. Cerantola, Y. et al. Perioperative nutrition in abdominal surgery: recommendations and reality. Gastroenterol. Res. Pract.2011
, 739347 (2011).
17. Healy, K. O. et al. Perioperative outcome and long term mortality for heart failure patients undergoing intermediate and high risk non-cardiac surgery: impact of left ventricular ejection fraction. Congest. Heart Fail.16
, 45-49 (2010).
18. Jain, U. et al. Predictors of readmission after inpatient plastic surgery. Arch. Plast. Surg.41
, 116-121 (2014).
19. Nguyen, G. C., Elnahas, A. & Jackson, T. D. The impact of preoperative steroid use on short-term outcomes following surgery for inflammatory bowel disease. J. Crohns Colitis8
, 1661-1667 (2014).
20. Fleshner, P. R. & Schoetz, D. J. in The ASCRS Textbook of Colon and Rectal Surgery
. pp. 567-583. Springer Science, Business Media, LLC, https://doi.org/10.1007/978-0-387-36374-5_41
21. Dozois, E. J. Proctocolectomy and Brooke ileostomy for chronic ulcerative colitis. Clin. Colon Rectal Surg.17
, 65-70 (2004).
22. Kaplan, G. G. et al. Impact of hospital volume on postoperative morbidity and mortality following a colectomy for ulcerative colitis. Gastroenterology134
23. Aziz, F. Return to operating room after lower extremity arterial bypass is an independent predictor for hospital readmission. J. Vasc. Surg.62
, 808 (2015).
24. Hardy, K. L. et al. The impact of operative time on complications after plastic surgery: a multivariate regression analysis of 1753 cases. Aesthetic Surg. J.34
, 614-622 (2014).
25. Merkow, R., Ju, M. & Chung, J. Underlying reasons associated with hospital readmission following surgery in the United States. J. Vasc. Surg.62
, 265 (2015).
26. Palmer, A. J. et al. Low molecular weight heparin and unfractionated heparin for prevention of thrombo-embolism in general surgery: a meta-analysis of randomised clinical trials. Haemostasis27
, 65-74 (1997).
27. Wald, H. L., Ma, A., Bratzler, D. W. & Kramer, A. M. Indwelling urinary catheter use in the postoperative period: analysis of the national surgical infection prevention project data. Arch. Surg.143
, 551-557 (2008).
28. Kaboli, P. J. et al. Associations between reduced hospital length of stay and 30-day readmission rate and mortality: 14-year experience in 129 veterans affairs hospitals. Ann. Intern. Med.157
, 837 (2015).
29. Bueno, H. et al. Trends in length of stay and short-term outcomes among Medicare patients hospitalized for heart failure: 1993-2008. JAMA303
, 2141-2147 (2010).
30. Coulam, R. F. & Gaumer, G. L. Medicare's prospective payment system: a critical appraisal. Health Care Financ. Rev.1991
(Suppl), 45-77 (1992).
31. Ljungqvist, O., Scott, M. & Fearon, K. C. Enhanced recovery after surgery: a review. JAMA Surg.152
, 292-298 (2017).
32. Kumar, A. et al. Inflammatory bowel disease: perioperative pharmacological considerations. Mayo Clin. Proc.86
, 748-757 (2011).