Your SlideShare is downloading. ×
98 DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014)
BACKGROUND: Colon resections are associated with
substantial risk fo...
DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 99
may also be subject to reduced reimbursement. Hospitals
with higher-...
KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION100
known bleeding disorder (including congenital clotting
disorde...
DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 101
vs 3.1%; p < 0.0001), 5) septic shock (4.8% vs 1.0%; p
< 0.0001), 6...
KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION102
readmissions are indeed a problem. Although the 30-day
mortali...
DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 103
demonstrating the superiority of ERPs over the conven-
tional posto...
KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION104
10. Fink AS, Campbell DA Jr, Mentzer RM Jr, et al. The Nationa...
Upcoming SlideShare
Loading in...5
×

Risk factors for readmission after elective colectomy postoperative complications are more important than patient and operative factors

129

Published on

Published in: Health & Medicine
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
129
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
2
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Transcript of "Risk factors for readmission after elective colectomy postoperative complications are more important than patient and operative factors"

  1. 1. 98 DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) BACKGROUND: Colon resections are associated with substantial risk for morbidity and readmissions, and these have become markers for quality of care. OBJECTIVE: The purpose of this study was to determine risk factors for readmissions after elective colectomies to improve patient care and better understand the complex issues associated with readmissions. DESIGN: This was an analysis of the prospective, statewide, multicenter Michigan Surgical Quality Collaborative database. SETTINGS: The analysis was conducted at academic and community medical centers in the state of Michigan. PATIENTS: Elective laparoscopic and open ileocolic and segmental colectomies from 2008 through 2010 were included. MAIN OUTCOME MEASURES: Univariate analysis and a multivariate logistic regression model were used to determine influence of patient characteristics, operative factors, and postoperative complications on the incidence of 30-day postoperative readmission. RESULTS: The readmission rate among 4013 cases was 7.3% (N = 293). On the basis of multivariate logistic regression, the top 3 significant risk factors associated with readmission were stroke (OR, 10.0 [95% CI, 2.70–37.0]; p = 0.001), venous thromboembolism (OR, 6.5 [95% CI, 3.7–11.3]; p < 0.0001), and organ-space surgical site infection (OR, 5.6 [95% CI, 3.4–9.4]; p < 0.0001). Important factors that contributed to readmission risk but were not found to be independent predictors of readmission included diabetes mellitus, preoperative steroids, smoking, cardiac comorbidities, age >80 years, anastomotic leaks, fascial dehiscence, sepsis, pneumonia, unplanned intubation, and length of stay. LIMITATIONS: The Michigan Surgical Quality Collaborative is a large database, and true causal relations are difficult to determine; reason for readmission is not recorded in the database. CONCLUSIONS: Postoperative complications account for the majority of risk factors behind readmissions after elective colectomy, whereas preoperative risk factors have less direct influence. Current strategies addressing readmission rates should focus on reducing preventable complications. KEY WORDS: Colon resection; Complications; Postoperative readmissions, Risk factors. C olectomies are frequently performed procedures with a high risk for complications and readmis- sions.1–4 Unplanned readmissions after colon and rectal operations can have serious consequences for pa- tients4 and have a significant economic impact for health- care providers. More recently, readmissions have gained widespread interest among insurance companies, media, and politicians as a quality indicator for medical and sur- gical outcomes. With the introduction of the Hospital Readmission Reduction Program by the Centers for Medi- care & Medicaid Services in 2012, 30-day readmissions Risk Factors for Readmission After Elective Colectomy: Postoperative Complications Are More Important Than Patient and Operative Factors Therese G. Kerwel, M.D.1 2 1 Samantha K. Hendren, M.D.3 2 1 1 Division of Colon and Rectal Surgery, Spectrum Health, Grand Rapids, Michigan 2 Department of Surgery, St. Joseph Mercy Health System, Ann Arbor, Michigan 3 Department of Surgery, University of Michigan, Ann Arbor, Michigan Dis Colon Rectum 2014; 57: 98–104 DOI: 10.1097/DCR.0000000000000007 © The ASCRS 2013 Financial Disclosure: None reported. Podium presentation at the meeting of The American Society of Colon and Rectal Surgeons, Phoenix, AZ, April 27 to May 1, 2013. Correspondence: Therese G. Kerwel, M.D., 25 Michigan St NE, Suite 2200, Grand Rapids, MI 49503. E-mail: theresegannon@hotmail.com ORIGINAL CONTRIBUTION
  2. 2. DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 99 may also be subject to reduced reimbursement. Hospitals with higher-than-predicted rates of readmissions may face financial penalties, and almost half of large teaching hos- pitals might be subjected to these cuts.5 Because of the enormous medical and financial im- pacts of unplanned readmissions, numerous studies have attempted to define risk factors for readmissions after colon and rectal surgery. Despite numerous studies, widespread disagreement remains in the medical literature regarding whether there are modifiable risk factors for readmission or whether it is largely unavoidable.6,7 None of those stud- ies were able to delineate definitive strategies for readmis- sion reduction,8 but all of them agree that the medical and financial costs associated with this are high. Identifying modifiable and/or avoidable factors in colon and rectal surgery would have the potential to improve patient out- comes and reduce health care-related costs. The aim of this study was to identify risk factors that are independently as- sociated with readmission within 30 days of elective colec- tomy by analyzing the prospective, multicenter Michigan Surgical Quality Collaborative (MSQC) database. MATERIALS AND METHODS The MSQC is a collaboration of 52 community and aca- demic hospitals throughout the state of Michigan with the goal to improve quality and patient outcomes.9 Sixty- two percent of the participating hospitals are community based without teaching activities.9 The MSQC database is similar to the rigorously maintained American College of Surgeons’ National Surgical Quality Improvement Pro- gram (NSQIP).10,11 Collected data include >200 variables and include patient demographic characteristics and co- morbidities, preoperative and intraoperative measures, and 30-day postoperative outcomes.11 The MSQC colecto- my project involves 24 of the 52 hospitals and additionally records procedure-specific perioperative data on patients undergoing partial colectomy. Data for the MSQC are meticulously recorded by NSQIP-certified clinical nurse reviewers at each participating hospital. In case of miss- ing or inconsistent data, the data abstractors communi- cate directly with the operating surgeon. Readmissions are tracked across all of the MSQC-participating hospitals and are not limited to the hospital where the patient un- derwent surgery. All of the cases of elective laparoscopic and open il- eocolic and segmental colectomies included as part of the MSQC colectomy project from June 2008 through Novem- ber 2010 were eligible for inclusion in our analysis. These operations were defined by Current Procedural Terminol- ogy codes 44140 and 44160 (open colectomy), as well as 44204 and 44205 (laparoscopic colectomy). The MSQC database excludes patients with age <18 years, pregnant women, trauma patients, and patients with ASA classes 5 and 6. Operations resulting in the creation of an ostomy, rectal resections, or low anastomoses are not included in the MSQC colectomy project. In addition, cases with com- plete obstruction or perforation, as well as any emergent cases, were also excluded from this study, because these are not elective and risk factors are not modifiable. To determine risk factors independently associated with readmission within 30 days of the index operation, patient characteristics, operative factors, and postopera- tive complications recorded in the MSQC database were first subjected to univariate analysis, determining the rela- tionship of each variable to 30-day readmission. This was followed by a multivariate logistic regression model, of both preoperative and postoperative variables, which in- cluded all of the variables with a p value of <0.10.Variables with significance on each multivariate logistic regression model were then entered into a final multivariate logistic regression model that combined preoperative and postop- erative variables to determine their independent influence on the incidence of 30-day postoperative readmission. RESULTS From June 2008 through November 2010, 4013 cases met inclusion criteria. Patient demographic characteristics are shown in Table 1. Their average age was 64.2±15.0 years with a mean BMI of 28.5±8.3. The readmission rate among these patients was 7.3% (N = 292). Table 2 lists the preoperative variables studied. Uni- variate analysis demonstrated a host of preoperative factors with associations to readmission: 1) ASA class 4 (p = 0.02), 2) being functionally dependent (either par- tial or total, p < 0.0001), 3) diabetes mellitus (p = 0.04), 4) hypertension (p = 0.003), 5) severe COPD (p = 0.02), 6) preoperative dyspnea (p = 0.001), 7) presence of an open wound (p = 0.011), 8) steroids (p < 0.0001), 9) TABLE 1. Demographics Entire cohort (N = 4013) Readmit (N = 292) No readmit (N = 3721) p N, % (n) 100 (4013) 7.28 (292) 92.72 (3721) Age (mean ± SD), y 64.2±15.0 65.0±16.0 64.2±15.0 0.38 Age ≥80 y, % (n) 16.7 (672) 20.9 (61) 16.4 (611) 0.06 Men, % (n) 47.7 (1914) 51.0 (149) 47.4 (1765) 0.25 BMI, mean ± SD 28.5±8.3 28.7±6.9 28.5±8.4 0.26 Black, % (n) 11.3 (452) 12.0 (35) 11.2 (417) 0.70
  3. 3. KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION100 known bleeding disorder (including congenital clotting disorders and chronic anticoagulation, p < 0.0001), and 10) anemia (hematocrit <30%, p = 0.04). The readmis- sion group was found to have, on average, more comor- bid conditions as defined by MSQC than the no-readmit group: 2.06±1.80 vs 1.57±1.50 (p = 0.0003). See the bottom of Table 2 for a definition of these comorbid con- ditions. A multiple logistic regression analysis of the pre- operative variables with a p value of ≤0.10 was performed to determine independent preoperative risk factors for readmission. The only significant variable resulting from this analysis was chronic preoperative steroid use (OR, 1.9 [95% CI, 1.14–3.19]; p = 0.014). Tables 3 and 4 show preoperative laboratory data and surgical variables studied. Factors on univariate analysis that were found to have an association with readmission are anemia (p = 0.04) and wound class 4 (p = 0.007). The readmission group also had more intraoperative blood loss than the no-readmit group (0.047). Laparoscopy showed a trend with decreased readmission rates, but this did not reach statistical significance. Wound class 4 was included in the analysis of preoperative variables and found to be an independent predictor of readmission (OR, 1.15; p = 0.0005) Table 5 shows the postoperative outcomes studied. The association between complications and readmission is striking from a statistical standpoint, with the major- ity of the complications being highly statistically signifi- cant for their association with readmission. Univariate analysis showed the following complications to be sig- nificant: 1) return to the operating room within 30 days (readmissions, 22.3%; no readmission, 3.6%; p < 0.0001), 2) anastomotic leak (12.8% vs 2.1%; p < 0.0001), 3) pro- longed ileus (21.4% vs 6.1%; p < 0.0001), 4) sepsis (14.7% TABLE 2. Preoperative variables and risk factors Entire cohort (N = 4013) Readmit (N = 292) No readmit (N = 3721) p Total risk factors, mean ± SDa 1.78±1.7 2.06±1.8 1.57±1.5 <0.0001 Smoker, % (n) 18.7 (744) 12.1 (35) 19.2 (709) 0.004 Alcohol use, % (n) 3.2 (127) 2.1 (6) 3.3 (121) 0.38 Functionally dependent, % (n) 4.8 (193) 10.3 (30) 4.4 (163) <0.0001 ASA, % (n) 1 2.7 (108) 2.4 (7) 2.7 (101) 1.00 2 50.3 (2020) 46.2 (135) 50.7 (1885) 0.13 3 42.6 (1708) 44.2 (129) 42.4 (1579) 0.58 4 4.3 (174) 7.2 (21) 4.1 (153) 0.02 Hypertension, % (n) 56.4 (2245) 63.4 (184) 55.8 (2061) 0.003 Diabetes mellitus, % (n) 17.7 (709) 21.6 (63) 17.4 (646) 0.04 Dyspnea, % (n) 16.3 (649) 21.4 (62) 15.9 (587) 0.001 Cardiac disease, % (n) 13.9 (552) 17.2 (50) 13.6 (502) 0.09 Severe COPD, % (n) 5.8 (233) 8.6 (25) 5.6 (208) 0.016 CVA, % (n) 5.0 (200) 5.9 (17) 5.0 (183) 0.40 Bleeding disorder, % (n) 4.2 (168) 7.9 (23) 3.9 (145) <0.0001 TIA, % (n) 4.0 (160) 4.8 (14) 4.0 (146) 0.44 Steroid use, % (n) 3.7 (149) 7.6 (22) 3.4 (127) <0.0001 10% weight loss, % (n) 3.3 (131) 3.1 (9) 3.3 (122) 1.00 Disseminated cancer, % (n) 2.6 (102) 2.4 (7) 2.6 (95) 1.00 Open wound, % (n) 1.1 (42) 2.4 (7) 0.9 (35) 0.011 DNR, % (n) 1.0 (38) 1.4 (4) 0.9 (34) 0.35 CHF within 30 days, % (n) 1.0 (41) 1.7 (5) 1.0 (36) 0.12 PVD, % (n) 1.0 (40) 1.7 (5) 0.9 (35/3694) 0.21 Dialysis dependence, % (n) 0.8 (31) 0.7 (2) 0.8 (29/3694) 1.00 Ascites, % (n) 0.4 (16) 0.0 (0) 0.4 (16/3694) 0.62 Delirium, % (n) 0.2 (7) 0.3 (1) 0.2 (6/3694) 0.41 DNR = do not resuscitate; COPD = chronic obstructive pulmonary disease; CHF = congestive heart failure; PVD = peripheral vascular disease; TIA = transient ischemic attack; CVA = cerebrovascular accident. Bold p values indicate significance. a Subjects receive 1 point for each of the following preoperative risk factors as defined by the Michigan Surgical Quality Collaborative: diabetes mellitus, preoperative sepsis, current smoker, >2 alcoholic drinks per day within 2 weeks, DNR status, ventila- tor dependence within 48 hours, severe COPD, current pneumonia, ascites within 30 days, esophageal varices, CHF within 30 days, history of myocardial infarction in the past 6 months, previous percutaneous coronary intervention/percutaneous transluminal coronary angioplasty, previous cardiac surgery, angina within 30 days, hypertension requiring medication, history of revasculariza- tion or amputation for PVD, rest pain/gangrene, acute renal failure within 24 hours, dialysis dependence, impaired sensorium within 48 hours, coma, hemiplegia/hemiparesis, history of TIAs, CVA with or without residual neurologic deficit, tumor involving central nervous system, paraplegia/paraparesis, quadriplegia/quadriparesis, disseminated cancer, open wound, steroid use for chronic condition, >10% loss of body weight in the last 6 months, ≥4 packed red blood cells preoperatively within 72 hours, chemotherapy or radiotherapy within 90 days, sepsis within 48 hours, pregnancy, or previous operation within 30 days.
  4. 4. DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 101 vs 3.1%; p < 0.0001), 5) septic shock (4.8% vs 1.0%; p < 0.0001), 6) pneumonia (5.5% vs 2.1%; p < 0.0001), 7) unplanned intubation (5.8% vs 2.4%; p < 0.0001), 8) or- gan space surgical site infection ((SSI) 16.4% vs 1.6%; p < 0.0001), 9) venous thromboembolism (9.2% vs 1.1%; p < 0.0001), 10) Clostridium difficile colitis (5.5% vs 1.1%; p < 0.0001), 11) mechanical bowel obstruction (6.6% vs 0.7%; p < 0.0001), 12) wound dehiscence (2.4% vs 0.8%; p = 0.010), 13) acute myocardial infarction (3.1% vs 0.5%; p < 0.0001), 14) cardiac arrest requiring cardiopulmonary resuscitation (1.7% vs 0.6%; p = 0.024), 15) stroke (1.7% vs 0.2%; p < 0.0001), 16) progressive renal insufficiency (4.8% vs 0.5%; p < 0.0001), 17) superficial SSI (14.0% vs 5.1%; p < 0.0001), and 18) urinary tract infection (6.5% vs 2.2%; p < 0.0001). The postoperative variables with a p value of ≤0.10 were evaluated in a multiple logistic regression analysis to determine which variables were independent risk fac- tors. These are reoperation within 30 days, organ space SSI, venous thromboembolism, Clostridium difficile coli- tis, acute myocardial infarction, stroke, mechanical bowel obstruction, prolonged ileus, urinary tract infection, and superficial SSI. These postoperative variables were then analyzed with preoperative steroid use in a combined multilogistic regression analysis. This then identified risk factors that independently predict readmission (Table 6): 1) postoperative stroke (OR, 10.01; p = 0.001), 2) postop- erative venous thromboembolism (OR, 6.51; p < 0.0001), 3) organ space SSI (OR, 5.6; p < 0.0001), 4) postopera- tive progressive renal insufficiency (OR, 3.95; p = 0.001), 5) postoperative Clostridium difficile colitis (OR, 3.61; p < 0.0001), 6) reoperation within 30 days (OR, 3.12; p < 0.0001), 7) postoperative myocardial infarction (OR, 2.92; p = 0.03), 8) postoperative mechanical bowel obstruction (OR, 2.92; p = <0.0001), 9) superficial SSI (OR, 2.78; p < 0.0001), 10) prolonged ileus (OR, 2.23; p < 0.0001), and 11) urinary tract infection (OR, 1.9; p = 0.008). Preopera- tive steroid use was no longer a significant risk factor when analyzed with postoperative complications. DISCUSSION Major bowel resections have a high rate of patient re- admission, with the majority (82%) occurring in the first 30 days postoperatively.12,13 Thirty-day readmission rates have become a marker of surgical quality and pos- sible grounds for reimbursement penalties. As Centers for Medicare & Medicaid Services attempts to involve read- missions as a component of the total cost of an episode of care, it is crucial that surgeons identify risk factors associ- ated with readmissions.14 This analysis of a large series of elective colectomies in multiple hospitals across the state of Michigan demonstrated a readmission rate of 7.3%. Although these results compare very favorably with fre- quently reported 30-day readmission rates between 11.0% and 13.7% after colorectal surgery,1,2,4 they confirm that TABLE 4. Surgical variables Cohort Readmit No readmit p Type of surgery Partial colectomy, % (n) 32.3 (1295) 36.6 (107) 31.9 (1188) 0.2 Right hemicolectomy, % (n) 19.9 (799) 19.9 (58) 19.9 (741) 1.00 Laparoscopic partial colectomy, % (n) 33.7 (1353) 30.1 (88) 34.0 (1265) 0.20 Laparoscopic right hemicolectomy, % (n) 14.1 (565) 13.3 (39) 14.1 (526) 0.79 Laparoscopy, % (n) 47.8 (1918) 43.5 (127) 48.1 (1791) 0.13 Intraoperative blood loss (mean ± SD), mL 137±220 156±180 135±171 0.047 Wound class, % (n) 2 86.6 (3474) 81.2 (237) 87.0 (3237) 0.007 3 9.9 (398) 12.3 (36) 9.7 (362) 0.16 4 3.5 (139) 6.5 (19) 3.2 (120) 0.0069 Bold p values indicate significance. TABLE 3. Selected laboratory values Cohort Readmit No readmit p Anemia (Hct <30%), % (n) 9.4 (362) 12.5 (35) 9.2 (327) 0.036 Mean Hct (mean ± SD), % 38.0±5.8 37.2±5.9 38.1±5.8 0.013 Preoperative glucose (mean ± SD), mg/dL 111±38 114±37 111±38 0.218 Intraoperative glucose, mean ± SD 158±58 166±51 157±59 0.47 Glucose POD1, mean ± SD 143±44 147±52 143±43 0.17 Glucose POD2, mean ± SD 126±37 128±38 126±37 0.46 Hct = hematocrit; POD = postoperative day. Bold p values indicate significance.
  5. 5. KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION102 readmissions are indeed a problem. Although the 30-day mortality rate in patients who were readmitted was not significantly higher than in patients who did not return to the hospital (2.1% vs 1.7%; p = 0.64), other studies have shown worse long-term outcomes at 1 year, especially among cancer patients.1,4 In our multivariate analysis, we demonstrated that postoperative complications accounted for the majority of risk factors that were independently associated with 30- day readmissions. Although not all postoperative compli- cations are avoidable, this factor may be more amenable to strategies aimed to reduce preventable complications than patient factors and comorbidities, because many patient factors cannot be modified in the preoperative period before colon resection. The strong association between readmissions and postoperative complications has been shown before in other studies.1,2,15,16 Of interest, the article by Kassin et al16 reached the same conclusions that we have using a similar analysis with similar methodology for gen- eral surgery patients. The complications with the highest incidence in our cohort were prolonged ileus (7.2%), superficial SSI (5.7%), sepsis (4.0%), anastomotic leak (3.3%), and organ space SSI (2.6%). All of the other complications related to read- missions had an incidence of well below 3%. These data suggest that an effort to reduce those complications might reduce readmission rates. A reduction in prolonged ileus has been shown with narcotic-sparing enhanced recov- ery pathways (ERPs) and use of minimally invasive tech- niques.17–19 As more high-quality studies become available TABLE 5. Postoperative outcomes Cohort Readmits No readmits p In-hospital death, % (n) 1.7 (70) 2.1 (6) 1.7 (64) 0.64 Prolonged ileus, % (n) 7.2 (285) 21.4 (62) 6.1 (223) <0.0001 SSI–superficial, % (n) 5.7 (230) 14.0 (41) 5.1 (189) <0.0001 Return to OR within 30 days, % (n) 4.9 (197) 22.3 (66) 3.6 (132) <0.0001 Sepsis, % (n) 4.0 (159) 14.7 (43) 3.1 (116) <0.0001 Anastomotic leak, % (n) 3.3 (113) 12.8 (37) 2.1 (76) <0.0001 Unplanned intubation, % (n) 2.6 (105) 5.8 (17) 2.4 (88) <0.0001 SSI–organ space, % (n) 2.6 (106) 16.4 (48) 1.6 (58) <0.0001 UTI, % (n) 2.5 (100) 6.5 (19) 2.2 (81) <0.0001 Pneumonia, % (n) 2.4 (95) 5.5 (16) 2.1 (79) <0.0001 VTE, % (n) 1.7 (69) 9.2 (27) 1.1 (42) <0.0001 Clostridium difficile colitis, % (n) 1.4 (58) 5.5 (16) 1.1 (42) <0.0001 Septic shock, % (n) 1.2 (50) 4.8 (14) 1.0 (36) <0.0001 Mechanical obstruction, % (n) 1.1 (43) 6.6 (19) 0.7 (24) <0.0001 Wound dehiscence, % (n) 1.0 (39) 2.4 (7) 0.8 (32) 0.010 Progressive renal insufficiency, % (n) 0.8 (33) 4.8 (14) 0.5 (19) <0.0001 SSI–deep space, % (n) 0.7 (30) 1.0 (3) 0.7 (27) 0.48 Acute myocardial infarction, % (n) 0.7 (29) 3.1 (9) 0.5 (20) <0.0001 Cardiac arrest, % (n) 0.7 (28) 1.7 (5) 0.6 (23) 0.024 Acute renal failure, % (n) 0.6 (23) 1.0 (3) 0.5 (20) 0.23 Stroke, % (n) 0.3 (11) 1.7 (5) 0.2 (6) <0.0001 Median length of stay, days 5 (0–82) 6 (1–58) 5 (0–82) SSI = surgical site infection; OR = operating room; UTI = urinary tract infection; VTE = venous thromboembolism. Bold p values indicate significance. TABLE 6. Risk factors for readmission based on multivariate analysis OR 95% CI, lower 95% CI, upper p Postoperative CVA 10.0 2.70 37.0 0.001 Postoperative VTE 6.51 3.75 11.31 <0.0001 Organ space SSI 5.63 3.38 9.37 <0.0001 Postoperative progressive renal insufficiency 3.95 1.74 8.94 0.001 Postoperative Clostridium difficile colitis 3.61 1.81 7.18 <0.0001 Postoperative reoperation 3.12 2.03 4.79 <0.0001 Superficial SSI 2.78 1.87 4.16 <0.0001 Postoperative mechanical bowel obstruction 2.92 1.35 6.33 0.007 Postoperative acute myocardial infarction 2.92 1.12 7.57 0.028 Prolonged ileus 2.23 1.53 3.24 <0.0001 Postoperative UTI 1.89 1.05 3.38 0.033 CVA = cerebrovascular accident; SSI = surgical site infection; UTI = urinary tract infection; VTE = venous thromboembolism. Bold p values indicate significance.
  6. 6. DISEASES OF THE COLON & RECTUM VOLUME 57: 1 (2014) 103 demonstrating the superiority of ERPs over the conven- tional postoperative management of colorectal patients, a continuing decrease in prolonged ileus may be seen. How- ever, reducing the rate of superficial SSIs has proven to be a difficult task despite tremendous effort on the part of physi- cians and hospitals. Although process measures such as the Surgical Care Improvement Program were introduced to reduce infectious complications after surgery and are now considered standard of care, the ability of these measures to actually improve outcomes is highly controversial.15,20,21 This leads to the more general question of how to ap- proach complications in colorectal surgery that are difficult to impact.Similar to superficial SSIs,anastomotic leaks have been studied extensively, but undisputed risk factors and strategies to avoid them, short of avoiding an anastomosis, remain elusive.Penalization for readmission related to some complications may, therefore, not be the right approach. Notably, several important patient-related factors, such as age,diabetes mellitus,and cardiac comorbidities were not found to be independent predictors of 30-day readmission, although they must contribute to some degree by increasing the risk for complications. In contrast to our study, Schnei- der et al1 found patient-related factors such as older age and severity of comorbidities to be important predictors, and Wick et al2 found an association between a longer initial length of stay (LOS (>7 days)) and readmission. Although we did not find that older age was a risk factor, our data do indicates that accumulating risk factors puts a patient at risk for readmission (OR, 1.22), although not to the degree that postoperative complications do (Table 6). The chain of events among preoperative risk factors, postoperative com- plications, and readmission is difficult to unravel. For many patient factors and complications, this degree of correlation between preoperative risk factors and postoperative com- plications will always remain unclear, and any analysis of a database that collects preoperative and postoperative data, such as the NSQIP, does not allow us to completely unravel this chain of events.This is a limitation that our study shares with other recent similar analyses1-3 on readmission, as well as most studies involving the NSQIP in general.1,16,22 The commonly used methodology of performing univariate fol- lowed by multivariate analysis aims to reduce confounding factors but cannot completely exclude them. Our results did not demonstrate an association be- tween LOS and readmission rate. This finding is particu- larly interesting in light of the increasing introduction of ERPs, which have been shown to lead to a reduction in LOS but have raised concerns about increased complica- tions and readmission rates. These findings are consistent with an increasing number of studies that confirm the safety of ERPs.3,8,17–19,23 When discussing these findings, several strengths and limitations of our study methodology have to be consid- ered. Our results are derived from a thoroughly main- tained multicenter database that adheres to strict data recording and quality standards. We analyzed a compre- hensive list of variables for their association with post- operative readmission, many of which would be difficult or impossible to assess in a randomized clinical trial. Be- cause of its inclusion of a wide variety of hospitals, from a rural community to large university hospitals, our patient population may be considered representative of a real- world patient care setting. Nevertheless, this study comes with the limitations associated with any database analy- sis, and true causal relations cannot be determined. The MSQC database does not record reasons for readmissions within 30 days, and, therefore, it is not possible to deter- mine whether some readmissions were planned. In addi- tion, our findings might be influenced by current surgical practices in the state of Michigan, which may limit their national generalizability. In summary, postoperative complications account for the majority of risk factors behind readmissions after elective colectomy, whereas preoperative, patient-related risk factors have less direct influence. Current strategies addressing readmission rates should focus on reducing preventable complications while still accepting that much may not be completely avoidable. In addition, the issue of readmission may need to be refocused on defining accept- able rates of complications instead of widespread policies aimed at penalizing all readmissions. REFERENCES 1. Schneider EB, Hyder O, Brooke BS, et al. Patient readmission and mortality after colorectal surgery for colon cancer: impact of length of stay relative to other clinical factors. J Am Coll Surg. 2012;214:390–398. 2. Wick EC, Shore AD, Hirose K, et al. Readmission rates and cost following colorectal surgery. Dis Colon Rectum. 2011;54:1475–1479. 3. Hendren S, Morris AM, Zhang W, Dimick J. Early discharge and hospital readmission after colectomy for cancer. Dis Colon Rec- tum. 2011;54:1362–1367. 4. Greenblatt DY, Weber SM, O’Connor ES, LoConte NK, Liou JI, Smith MA. Readmission after colectomy for cancer predicts one-year mortality. Ann Surg. 2010;251:659–669. 5. Joynt KE, Jha AK. Characteristics of hospitals receiving pen- alties under the Hospital Readmissions Reduction Program. JAMA. 2013;309:342–343. 6. Joynt KE, Jha AK. Thirty-day readmissions: truth and conse- quences. N Engl J Med. 2012;366:1366–1369. 7. van Walraven C, Bennett C, Jennings A, Austin PC, Forster AJ. Proportion of hospital readmissions deemed avoidable: a sys- tematic review. CMAJ. 2011;183:E391–E402. 8. O’Brien DP, Senagore A, Merlino J, Brady K, Delaney C. Predic- tors and outcome of readmission after laparoscopic intestinal surgery. World J Surg. 2007;31:2430–2435. 9. Campbell DA Jr, Kubus JJ, Henke PK, Hutton M, Englesbe MJ. The Michigan Surgical Quality Collaborative: a legacy of Shukri Khuri. Am J Surg. 2009;198(suppl):S49–S55.
  7. 7. KERWEL ET AL: RISK FACTORS FOR POSTOPERATIVE READMISSION104 10. Fink AS, Campbell DA Jr, Mentzer RM Jr, et al. The National Surgical Quality Improvement Program in non-veterans ad- ministration hospitals: initial demonstration of feasibility. Ann Surg. 2002;236:344–353. 11. Khuri SF, Daley J, Henderson W, et al. The Department of Vet- erans Affairs’ NSQIP: the first national, validated, outcome- based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg. 1998;228:491–507. 12. Azimuddin K, Rosen L, Reed JF 3rd, Stasik JJ, Riether RD, Khubchandani IT. Readmissions after colorectal surgery cannot be predicted. Dis Colon Rectum. 2001;44:942–946. 13. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med. 2009;360:1418–1428. 14. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skin- ner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45:1783–1795. 15. Berenguer CM, Ochsner MG Jr, Lord SA, Senkowski CK. Im- proving surgical site infections: using National Surgical Quality Improvement Program data to institute Surgical Care Improve- ment Project protocols in improving surgical outcomes. J Am Coll Surg. 2010;210:737–41, 741. 16. Kassin MT, Owen RM, Perez SD, et al. Risk factors for 30-day hospital readmission among general surgery patients. J Am Coll Surg. 2012;215:322–330. 17. Jorgensen H, Wetterslev J, Moiniche S, Dahl JB. Epidural lo- cal anaesthetics versus opioid-based analgesic regimens on postoperative gastrointestinal paralysis, PONV and pain after abdominal surgery. Cochrane Database Syst Rev. 2000;(4):CD001893. 18. Li MZ, Xiao LB, Wu WH, Yang SB, Li SZ. Meta-analysis of lapa- roscopic versus open colorectal surgery within fast-track peri- operative care. Dis Colon Rectum. 2012;55:821–827. 19. Vlug MS, Wind J, Hollmann MW, et al.; LAFA study group. Laparoscopy in combination with fast track multimodal man- agement is the best perioperative strategy in patients undergo- ing colonic surgery: a randomized clinical trial (LAFA-study). Ann Surg. 2011;254:868–875. 20. Garcia N, Fogel S, Baker C, Remine S, Jones J. Should com- pliance with the Surgical Care Improvement Project (SCIP) process measures determine Medicare and Medicaid reim- bursement rates? Am Surg. 2012;78:653–656. 21. Ingraham AM, Cohen ME, Bilimoria KY, et al. Associa- tion of surgical care improvement project infection-related process measure compliance with risk-adjusted outcomes: implications for quality measurement. J Am Coll Surg. 2010;211:705–714. 22. Schneider EB, Hyder O, Wolfgang CL, et al. Patient readmission and mortality after surgery for hepato-pancreato-biliary malig- nancies. J Am Coll Surg. 2012;215:607–615. 23. KolozsvariNO,CaprettiG,KanevaP,etal.Impactof anenhanced recovery program on short-term outcomes after scheduled lap- aroscopic colon resection. Surg Endosc. 2013;27:133–138.

×