Unplanned Return to Operating Room After Lower Extremity Arterial Bypass is an Independent Predictor for Hospital Readmission

Abstract

Objectives

Hospital readmissions after surgical operations are considered as serious complications and impact health care associated costs. Centers for Medicare and Medicaid services (CMS) strongly encourage identification and ramification of factors associated with hospital readmissions after operations. Despite advances in endovascular surgery, lower extremity arterial bypass remains the gold standard treatment for severe, symptomatic Peripheral Arterial Disease (PAD). The purpose of this study is to retrospectively review the factors associated with hospital readmission after lower extremity bypass surgery.

Methods

The 2013 lower extremity revascularization–targeted American College of Surgeons (ACSNSQIP) database and generalized 2013 general and vascular surgery ACS-NSQIP PUF were used for this study. Patient, diagnosis, and procedure characteristics of patients undergoing lower extremity bypass surgery were assessed. Multivariate logistic regression analysis was used to determine independent risk factors for hospital readmission within 30 days after surgery.

Results

A total of 2646 patients (Males 65%, , Females 35%) were identified in NSQIP database, who underwent lower extremity open revascularization during the year 2013. Indications for operations included: tissue loss (39%), rest pain (32%) and severe claudication (25%). Pre-operative ABIs were: 0.4–0.9 (32%) and <0.4 (16.5%). A total of 425 patients (16%) were readmitted within 30 days of index operation. Risk factors, associated with readmission included: wound complication (OR 8.54, CI 6.68–10.92, p<0.001), need for re-operation (OR 5.95, CI 4.45–7.97, p <0.001), post-operative MI (OR 2.19, CI 1.25–3.83, p=0.006), wound dehiscence (OR 8.45, CI 4.54–15.71, p<0.001), organ/space SSI (OR 7.62, CI 2.89–20.14, p<0.001), post-operative pneumonia (OR 2.66, CI 1.28–5.52, p=0.009), progressive renal insufficiency (OR 4.12, CI 1.52–11.11, p =0.005), superficial surgical site infection (OR 7.37, CI 5.31–10.23, p <0.001), urinary tract infection (OR 2.67, CI 1.42–5.01, p 0.002) and deep wound infection (OR 14.0, CI 7.62–24.80, p<0.001)

Conclusions

Readmission after lower extremity bypass surgery is a serious complication. Various factors put a patient at a high risk for readmission. Return to operating room, wound infection, amputation, DVT and major re-intervention on bypass are independent risk factors for hospital readmission. Return to operating room is associated with a 5.95 fold increase in hospital readmission.

Introduction

With decreasing lengths of hospital stay, incidence of early readmissions to hospitals has been increasing¹. Section 3025 of Patient Protection and Affordable Care Act² puts a greater responsibility over the hospitals for early readmissions. Majority of the surgical readmissions are related to post-operative complications³. Since the public release of a report by Institute of Medicine in 1999⁴ on medical errors, there has been an increasing public focus⁵ on recognizing and preventing potentially preventable adverse events. It is estimated that 48% to 66% of all hospital adverse events are related to surgery and more than half are deemed preventable^6–9. Unlike other post-operative complications, unplanned reoperations can be considered nondiscretionary and major events. Any unplanned return to operating room is tracked by administrative databases and by American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP). Patients with peripheral arterial disease represent the most severe form of atherosclerotic disease and hospital readmission after lower extremity open revascularization can be multifactorial. In an era of increasing health care scrutiny, it is crucial to identify potentially preventable causes of readmission. The purpose of this study is to identify the risk factors associated with increased risk of hospital re-admission and to analyze which of these factors can be considered preventable. This has implications for patient care quality and financial reimbursements in future.

Methods

Data set:

The American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) Participant Use File (PUF)¹⁰ is a de-identified data set, generated and operated by the ACS. The dataset is compliant with Health Insurance Portability and Accountability Act (HIPAA). It has more than 250 participant academic and community United States hospitals. Methods used to extract data from NSQIP database have been described previously^11–15. Data is collected by using a systematic sampling method. Surgical operations are divided into 8-day cycles. At each NSQIP site, the first forty operations performed within each 8-day that meet program inclusion criteria are entered in the database. NSQIP program limits the number of cases per cycle for certain higher volume and lower risk surgeries in order to ensure heterogeneity. At each ACS-NSQIP site, a trained clinical nurse is assigned for data collection. Outcomes have been shown to be highly reliable with less than 1.5% variable disagreements during annual audits¹³. To ensure complete follow-up, patients with incomplete 30-day outcomes are excluded from the database. Since there are no patient identifiers in NSQIP database, no Institutional Review Board (IRB) approval or patients’ consent was required.

Patients:

All patients who underwent any lower extremity open revascularization procedure during the year 2013, using Procedure Targeted Participant User File¹⁰ from NSQIP database. Using unique case identification numbers, this file was merged to the main ACS NSQIP adult Participant Use Data File. Any patient who presented with revascularization of bilateral limbs in the same calendar year was deleted from the data set.

Outcomes:

Primary outcome was re-admission within 30 days after surgery. Basic demographic data were analyzed including age, gender, race, age range and body mass index range. Several peri-operative variables were analyzed: operative times, length of hospital stay, type of operation, symptoms, high risk physiologic factors, high risk anatomic factors pre-operative use of aspirin, pre-operative use of beta-blockers, pre-operative use of statins, need for amputation, significant post-operative bleeding, post-operative myocardial infarction, post-operative stroke, untreated loss of patency, wound infection, pre-operative albumin, number of days from hospital admission to operation, type of anesthetic, American Society of Anesthesiology (ASA) classification, diabetes mellitus, end-stage renal disease, emergency operation, congestive heart failure, chronic obstructive pulmonary disease (COPD), hypertension, post-operative renal failure, need for re-operation, history of smoking, surgeons’ specialty, need for blood transfusion, transfer status, urinary tract infection, wound classification, cardiac arrest, wound disruption, superficial and deep wound infection, pneumonia and need for re-intubation (Table I)

Table I.

Factors associated with 30-day readmission after lower extremity open revascularization procedures

	No Readmission*	Readmission*
Variable	(N=2221)	(N=425)	OR (95% CI)+	P-value+
Pre-Operative factors
Age (years, 5-year increment)	67.6 ± 11.2	68.3 ± 11.4	1.03 (0.98, 1.08)	.27
Age Range (years)				.56
- <60	540 (84)	102 (16)	Reference
- 60–67	533 (84)	102 (16)	1.01 (0.75, 1.37)
- 68–75	590 (85)	102 (15)	0.92 (0.68, 1.23)
- ≥76	558 (82)	119 (18)	1.13 (0.85, 1.51)
Sex				.09
- Male	1453 (85)	260 (15)	Reference
- Female	768 (82)	165 (18)	1.20 (0.97, 1.49)
Race				.01
- Hispanic	103 (84)	20 (16)	1.11 (0.67, 1.82)
- Non-Hispanic Black	357 (79)	97 (21)	1.55 (1.20, 2.01)
- Non-Hispanic Other	32 (89)	4 (11)	0.71 (0.25, 2.03)
- Non-Hispanic White	1511 (85)	265 (15)	Reference
BMI Range (kg/m2)				.14
- <25	746 (84)	141 (16)	Reference
- 25 – <30	751 (86)	126 (14)	0.89 (0.68, 1.15)
- 30 – <40	598 (82)	133 (18)	1.18 (0.91, 1.53)
- ≥40	71 (80)	18 (20)	1.34 (0.78, 2.32)
Symptoms				<.01
- Asymptomatic	54 (98)	1 (22)	Reference
- Claudication	612 (90)	72 (10)	6.35 (0.87, 46.62)
- CLI - Rest Pain	701 (83)	143 (17)	11.02 (1.51, 80.28)
- CLI - Tissue Loss	827 (80)	205 (20)	13.39 (1.84, 97.33)
High Risk Physiologic Factors				.01
- Absent	1740 (85)	306 (15)	Reference
- Present	449 (80)	113 (20)	1.43 (1.13, 1.82)
Pre-op Aspirin				.72
- No	408 (83)	81 (17)	Reference
- Yes	1799 (84)	340 (16)	0.95 (0.73, 1.24)
Pre-op Beta-blockers				.01
- No	878 (87)	134 (13)	Reference
- Yes	1316 (82)	290 (18)	1.44 (1.16, 1.80)
Pre-op Statins				.76
- No	681 (84)	128 (16)	Reference
- Yes	1520 (84)	296 (16)	1.04 (0.83, 1.30)
Diabetes				.01
- None	1265 (86)	207 (14)	Reference
- Non-insulin dependent DM	384 (83)	79 (17)	1.26 (0.95, 1.67)
- Insulin dependent DM	572 (80)	139 (20)	1.49 (1.17, 1.88)
Dialysis Dependent				.31
- No	2102 (84)	397 (16)	Reference
- Yes	119 (81)	28 (19)	1.25 (0.81, 1.91)
Emergency Operation				.37
- No	2098 (84)	397 (16)	Reference
- Yes	122 (81)	28 (19)	1.21 (0.79, 1.85)
History of CHF				.04
- No	2167 (84)	407 (16)	Reference
- Yes	54 (75)	18 (25)	1.78 (1.03, 3.06)
History of COPD				.42
- No	1922 (84)	374 (16)	Reference
- Yes	299 (85)	51 (15)	0.88 (0.64, 1.20)
Hypertension				<.01
- No	398 (89)	48 (11)	Reference
- Yes	1823 (83)	377 (17)	1.72 (1.25, 2.36)
Smoking				.26
- No	1299 (83)	261 (17)	Reference
- Yes	922 (85)	164 (15)	0.89 (0.72, 1.10)
Transfer Status				.34
- From Acute Care hospital inpatient	90 (84)	17 (16)	1.02 (0.60, 1.73)
- Not transferred	1990 (84)	369 (16)	Reference
- Nursing home-chronic care-intermediate care	64 (78)	18 (22)	1.52 (0.89, 2.59)
- Outside ED	61 (78)	17 (22)	1.50 (0.87, 2.60)
- Transfer from other	16 (80)	4 (20)	1.35 (0.45, 4.06)
Pre-operative UTI				.85
- No	2212 (84)	423 (16)	Reference
- Yes	9 (82)	2 (18)	1.16 (0.25, 5.40)
Wound Classification				.34
- Clean	2100 (84)	396 (16)	Reference
- Clean/contaminated	50 (79)	13 (21)	1.38 (0.74, 2.56)
- Contaminated	27 (75)	9 (25)	1.77 (0.83, 3.79)
- Dirty/infected	44 (86)	7 (14)	0.84 (0.38, 1.89)
Open wound, present at the time of operation				<.01
- No	1486 (86)	244 (14)	Reference
- Yes	375 (80)	181 (20)	1.50 (1.21, 1.85)
Pre-op Albumin (mg/dl)	3.6 ± 0.7	3.5 ± 0.6	0.83 (0.67, 1.02)	.08
Days from Hospital admission to Operation (days)	1.9 ± 3.5	2.5 ± 3.9	1.04 (1.02, 1.07)	<.05
Intra-Operative Factors
Operative Time (minutes)				<.01
- 0–170	602 (89)	75 (11)	Reference
- 171–225	539 (85)	97 (15)	1.45 (1.05, 2.00)
- 226–300	528 (80)	136 (20)	2.07 (1.52, 2.81)
- >300	551 (82)	117 (18)	1.70 (1.25, 2.33)
Type of Operation				.72
- Femoral distal bypass with prosthetic/spliced vein/composite	220 (83)	45 (17)	1.11 (0.77, 1.61)
- Femoral distal bypass with single segment saphenous vein	404 (82)	87 (18)	1.17 (0.87, 1.57)
- Femoral endarterectomy	33 (94)	2 (6)	0.33 (0.08, 1.39)
- Femoropopliteal bypass with single segment sephanous vein	745 (84)	137 (16)	Reference
- Femoropopliteal bypass with prosthetic/spliced vein/composite	568 (84)	109 (16)	1.04 (0.79, 1.37)
- Other	52 (90)	6 (10)	0.63 (0.26, 1.49)
- Popliteal distal bypass with prosthetic/spliced vein/composite	46 (84)	9 (16)	1.06 (0.51, 2.22)
- Popliteal distal bypass with single segment sephanous vein	145 (83)	30 (17)	1.13 (0.73, 1.74)
- Profundoplasty	7 (100)	0 (0)	N/A
High Risk Anatomic Factors				.04
- Absent	1322 (83)	272 (17)	Reference
- Prior bypass involving same segment	529 (87)	77 (13)	0.71 (0.54, 0.93)
- Prior endovascular intervention involving same segment	370 (83)	76 (17)	1.00 (0.76, 1.32)
Type of Anesthetic				.94
- Epidural	24 (80)	6 (20)	1.30 (0.53, 3.20)
- General	2118 (84)	407 (16)	Reference
- MAC/IV Sedation	11 (100)	0 (0)	N/A
- Regional	4 (80)	1 (20)	1.30 (0.15, 11.67)
- Spinal	64 (85)	11 (15)	0.89 (0.47, 1.71)
ASA Class				.04
- 1–2 - No disturbance-Mild disturbance	121 (90)	14 (10)	Reference
- 3 - Severe disturbance	1641 (85)	296 (15)	1.56 (0.88, 2.75)
- 4–5 - Life threatening-Moribund	457 (80)	115 (20)	2.18 (1.21, 3.92)
Surgeon’s Specialty				.56
- Cardiac Surgery	4 (67)	2 (33)	2.59 (0.47, 14.19)
- General Surgery	34 (92)	3 (8)	0.46 (0.14, 1.50)
- Orthopedic Surgery	1 (100)	0 (0)	N/A
- Thoracic Surgery	52 (82)	9 (18)	0.90 (0.44, 1.83)
- Vascular Surgery	2130 (84)	411 (16)	Reference
Post-Operative Factors
Length of Hospital Stay (days)				<.01
- <7	1267 (86)	206 (14)	Reference
- 7–13	600 (80)	151 (20)	1.55 (1.23, 1.95)
- 14–20	212 (83)	43 (17)	1.25 (0.87, 1.79)
- 21–27	73 (79)	20 (21)	1.69 (1.01, 2.82)
- >28	69 (93)	5 (7)	0.45 (0.18, 1.12)
Amputation				<.01
- No	2173 (85)	380 (15)	Reference
- Yes	48 (52)	45 (48)	5.36 (3.52, 8.17)
Bleeding				<.01
- No	1891 (86)	299 (14)	Reference
- Yes	329 (72)	126 (28)	2.42 (1.91, 3.08)
Combined Ml/Stroke				<.01
- No	2164 (84)	400 (16)	Reference
- Yes	56 (69)	25 (31)	2.42 (1.49, 3.92)
Most Severe Procedural Outcome				<.01
- Clinically patent graft	1358 (87)	200 (13)	Reference
- Death	31 (89)	4 (11)	0.88 (0.31, 2.51)
- Graft Thrombosis with no intervention	16 (67)	8 (33)	3.40 (1.43, 8.04)
- Major Amputation	44 (54)	38 (46)	5.87 (3.71, 9.28)
- New bypass in the treated arterial segment	31 (77)	9 (23)	1.97 (0.93, 4.20)
- Other	29 (66)	15 (34)	3.51 (1.85, 6.67)
- Patent graft with stenosis	45 (82)	10 (18)	1.51 (0.75, 3.04)
- Patent graft, no stenosis	337 (83)	67 (17)	1.35 (1.00, 1.83)
- Revised graft with stenosis	12 (63)	7 (37)	3.96 (1.54, 10.18)
- Revised graft, no stenosis	23 (61)	15 (39)	4.43 (2.27, 8.63)
Untreated loss of patency				<.01
- No	2190 (84)	405 (16)	Reference
- Yes	30 (60)	20 (40)	3.61 (2.03, 6.41)
Wound Complication				<.01
- No	2041 (89)	243 (11)	Reference
- Yes	179 (57)	182 (43)	8.54 (6.68, 10.92)
Acute Renal Failure Post-Op				.82
- No	2203 (84)	422 (16)	Reference
- Yes	18 (86)	3 (14)	0.87 (0.26, 2.97)
Need for Re-Operation				<.01
- No	2009 (90)	216 (10)	Reference
- Yes	212 (50)	209 (50)	9.17 (7.24, 11.63)
Blood Transfusion (>4 unites in 72 hours before surgery)				.86
- No	2182 (84)	417 (16)	Reference
- Yes	39 (83)	8 (17)	1.07 (0.50, 2.31)
Cardiac arrest				.42
- No	2202 (84)	423 (16)	Reference
- Yes	19 (90)	2 (10)	0.55 (0.13, 2.36)
Post-op MI				.01
- No	2177 (84)	407 (16)	Reference
- Yes	44 (71)	18 (29)	2.19 (1.25, 3.83)
Wound disruption/dehiscence				<.01
- No	2204 (85)	399 (15)	Reference
- Yes	17 (39)	26 (61)	8.45 (4.54, 15.71)
Acute Renal Failure (post-op)				.77
- No	2208 (84)	423 (16)	Reference
- Yes	13 (87)	2 (13)	0.80 (0.18, 3.57)
Organ/Space SSI				<.01
- No	2214 (84)	415 (16)	Reference
- Yes	7 (41)	10 (59)	7.62 (2.89, 20.14)
Post-op Pneumonia				.01
- No	2199 (84)	414 (16)	Reference
- Yes	22 (67)	11 (33)	2.66 (1.28, 5.52)
Need for re-intubation				.34
- No	2188 (84)	416 (16)	Reference
- Yes	33 (79)	9 (21)	1.44 (0.68, 3.02)
Progressive Renal Insufficiency (Post-Op)				.05
- No	2212 (84)	418 (16)	Reference
- Yes	9 (56)	7 (44)	4.12 (1.52, 11.11)
Superficial surgical site infection				<.01
- No	2145 (86)	337 (14)	Reference
- Yes	76 (46)	88 (54)	7.37 (5.31, 10.23)
Post-Op UTI				.01
- No	2191 (84)	410 (16)	Reference
- Yes	30 (67)	15 (33)	2.67 (1.43, 5.01)
Deep Wound Infection				<.01
- No	2206 (85)	388 (15)	Reference
- Yes	15 (29)	37 (71)	14.0 (7.62, 25.80)

^*N (%) or Mean ± SD

⁺Odds ratios and p-values from logistic regression

Definition:

Only unplanned readmissions were included in this study. It was defined as any unplanned readmission (to the same or another hospital) for a post-operative occurrence likely related to the principal surgical procedure within 30 days of the procedure. Re-operation is defined as “an unplanned return to the operating room for a surgical procedure related to either the index or concurrent procedure performed”. This return must be within the 30-day postoperative period. The return to the operating room may occur at any hospital or surgical facility (i.e. index hospital or at an outside hospital). This definition is not meant to capture patients who go back to the operating room within 30 days for a follow-up procedure based on the pathology results from the index or concurrent procedure.” (Please see Appendix for NSQIP definitions)

Statistical Analysis:

All variables were initially summarized with frequencies and percentages or means, medians, and standard deviations. Logistic regression was used to determine any bivariate associations of independent variables with 30-day readmission. Odds ratios were used to quantify the magnitude and direction of any significant associations. The significant (p<0.05) independent variables from the bivariate analysis were then used in a process of stepwise selection to find the group of variables collectively that were most significantly associated with 30-day readmission in a multivariable logistic regression model. With so many variables and a large sample size, a more stringent entry criteria of p<0.05 and a stay criteria of p<0.05 were used for the stepwise process of variable selection to be more conservative. Forward and backward selection methods were also employed to check for other potential models, but the three approaches resulted in similar reduced models. The fit of the final model was checked using the Hosmer and Lemeshow goodness-of-fit test (p=0.1140). The c-statistic (c=0.820) was used to estimate the prediction strength of the final model. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

Results

Demographics and preoperative comorbidities:

A total of 2,646 patients (65% Males, 35% Females) underwent lower extremity revascularization operations in the year 2013. Mean age was 67.7 (± 11.3) years. Among these patients, 425 (16%) were readmitted to the hospital within 30 days after surgery. About 40% of all re-admissions were within two weeks after the discharge (Figure 1).

Figure 1. Readmission within 30 days after lower extremity bypass surgeries.

30 Day readmission rates among patients undergoing lower extremity bypass surgeries were as follows: 11% within first week after discharge, 40% within 2 weeks after discharge, 69% within 3 weeks after discharge and 96% within 4 weeks after discharge from the hospital.

Comparing variables between no readmission and readmission groups:

Patients were divided into two groups: no readmission (N=2221) and readmission (N=425) groups. The following factors were found to have no significant difference between these two groups: gender, age range, body mass index range, type of operation, high risk anatomic factors, pre-operative us of aspirin/beta-blockers/statins, pre-operative albumin, type of anesthetic, pre-operative dialysis dependency, emergency operation, history of COPD, post-operative renal failure, smoking, need for blood transfusions, transfer status, pre-operative urinary tract infection, wound classification, cardiac arrest, post-operative acute renal failure and need for re-intubation (Table I).

Bivariate analysis:

The following factors were found to be significantly associated with 30-day readmissions in bivariate analysis using logistic regression: race (non-Hispanic black versus non-Hispanic white (OR 1.55, CI 1.20–2.01, p-value = 0.009), pre-operative symptoms (Critical limb ischemia with tissue loss versus asymptomatic (OR 13.39, CI 1.84–97.33, p<0.001) (Critical limb ischemia with rest pain versus asymptomatic (OR 11.02, CI 1.51–80.28, p <0.001), presence of high risk physiologic factors (OR 1.43, CI 1.13–1.82, p=0.003), pre-op beta-blockers (OR 1.44, CI 1.16–1.80, p=0.001), diabetes mellitus (OR 1.49, CI 1.17–1.88, p=0.004), congestive heart failure (OR 1.78, CI 1.03–3.06, p = 0.039), hypertension (OR 1.72, CI 1.25–2.36, p <0.001), open wounds at the time of operation (OR 1.50, CI 1.21–1.85, p <0.001), operative time (225–300 minutes versus <170 minutes: OR 2.07, CI 1.52–2.81, p <0.001), American society for anesthesiology score (4–5 vs. 1–2: OR 2.18, CI 1.21–3.92, p=0.004, length of hospital stay (7–13 days vs. <7 days: OR 1.55, CI 1.23–1.95, p<0.001), amputation (OR 5.36, CI 3.52–8.17, p <0.001), bleeding (OR 2.42, CI 1.91–3.08, p<0.001), most severe procedural outcome (major amputation vs. clinically patent graft: OR 5.87, CI 3.71–9.28, p <0.001), untreated loss of patency (OR 3.61, CI 2.03–6.41, p <0.001), wound complication (OR 8.54, CI 6.68–10.92, p<0.001), need for re-operation (OR 9.17, CI 7.24–11.63, p <0.001), post-operative MI (OR 2.19, CI 1.25–3.83, p=0.006), wound dehiscence (OR 8.45, CI 4.54–15.71, p<0.001), organ/space SSI (OR 7.62, CI 2.89–20.14, p<0.001), post-operative pneumonia (OR 2.66, CI 1.28–5.52, p=0.009), progressive renal insufficiency (OR 4.12, CI 1.52–11.11, p =0.005), superficial surgical site infection (OR 7.37, CI 5.31–10.23, p <0.001), urinary tract infection (OR 2.67, CI 1.42–5.01, p 0.002) and deep wound infection (OR 14.0, CI 7.62–24.80, p<0.001) (Table I).

Reasons for readmission:

In majority of the patients (41.4%), reason for readmission could not be determined. Reason for readmission was determined not related to principal procedure in 22.8% of the patients. Among the readmissions in which the cause of readmission could be determined, the most common causes of readmission were: superficial skin infection (15.3%), deep incisional surgical site infection (7.1%), wound disruption (3.1%) and sepsis (1.7%) (Table II).

Table II.

Reasons for readmission

Reason for Readmission	Number of Patients	Percentage
Could not be determined	176	41.4%
Not related to principal procedure	97	22.8%
Superficial Skin Infection	65	15.3%
Deep Incisional SSI	30	7.1%
Wound disruption	13	3.1%
Unknown	18	4.2%
Sepsis	7	1.7%
Organ/Space SSI	3	0.7%
Pneumonia	3	0.7%
Pulmonary Embolism	3	0.7%
UTI	3	0.7%
Bleeding Requiring transfusion	2	0.5%
Progressive Renal Insufficiency	2	0.5%
CVA	1	0.2%
DVT Requiring Therapy	1	0.2%
MI	1	0.2%
Total	425	100.%

Multivariable analysis:

Stringent entry and stay criteria of p<0.05 were used for the stepwise process of variable selection to determine the best multivariable logistic regression model that included the factors most significantly associated with 30-day readmissions. The following factors were found to have significant associations with readmission: an operative time of 225300 minutes (OR 1.75, CI 1.19–2.57, p=0.004), physiologic high risk (OR 1.45, CI 1.07–1.96, p=0.016, need for amputation (OR 2.07, CI 1.24–3.47, p=0.006), bleeding (OR 1.51, CI 1.112.05, p 0.008), wound infection (OR 6.21, CI 4.65–8.30, p<0.001), re-operation (OR 5.95, CI 4.45–7.97, p<0.001) and progressive renal insufficiency (OR 3.99, CI 1.12–14.21, p = 0.033) (Table III).

Table III.

Multivariable model

Risk Factor	Adjusted Odd’s Ratio (95% CI)	P-Value
Operative Time 225–300 minutes	1.8 (1.19 – 2.57)	<.05
Physiological High Risk	1.5 (1.07 – 1.96)	<.05
Need for Amputation	2.1 (1.24 – 3.47)	<.01
Bleeding	1.5 (1.11 – 2.05)	<.01
Wound infection	6.2 (4.65 – 8.30)	<.01
Re-Operation	5.9 (4.45 – 7.97)	<.01
Renal Insufficiency	4 (1.12 – 14.21)	<.05

Predicted probability of readmission:

The probability of readmission was calculated for all of the factors identified to be significant in the multivariable analysis, either alone or combined. The probability of readmission was 6.5% for patients with high physiologic risk, 6.8% for patients who suffered significant bleeding, 7.8% for patients whose operative time was more than 225 minutes, 9.1% for patients who needed amputation, 16.1% for patients who developed progressive renal insufficiency in the post-operative period, 22.3% for those who needed reoperation, 23.1% for patients with wound infection, and 98.3% for patients who had all of these seven factors present (Table IV).

Table IV.

Predicted probability of readmission for risk factors from multivariable model

Physiologic High Risk	Bleeding	Operative Time (225–300 minutes)	Need for Amputation	Renal Insufficiency	Reoperation	Wound Infection	Probability of Readmission
+	−	−	−	−	−	−	6.5%
−	+	−	−	−	−	−	6.8%
−	−	+	−	−	−	−	7.8%
−	−	−	+	−	−	−	9.1%
−	−	−	−	+	−	−	16.1%
−	−	−	−	−	+	−	22.3%
−	-	−	−	−	−	+	23.1%
+	+	+	+	+	+	+	98.3%

^*Operative time is 0–170 minutes for all other calculations of the probability of readmission

Discussion

Early readmission to hospital after being discharged has always being considered as a reflection of suboptimal care provided during the hospital stay. Introduction of Hospital Prospective Payment System (PPS) in 1980s encouraged shorter hospital stays to reduce acute care costs. In the time period between 1993 and 2006, hospital length of stay for Medicare heart failure patients decreased by 25%¹. Associated with this decrease in length of hospital stay was 26% increase in 30-day hospital re-admission. With recent changes in the health care reimbursements, early readmission is increasingly becoming a focus of attention. A recent study analyzed the health care trends among Medicare patient population over a 27-year time period between 1976 and 2003 and found that the incidence of hospital readmission within 60 days increased from 23% to 31%¹⁶. In 2009, the centers for Medicare and Medicaid started publishing data for 30-day readmission data for certain medical diagnoses. In 2010, the Patient Protection and Affordable Care Act became the law. Section 3025 of this law holds the hospitals responsible for all patients who are readmitted within 30 days of discharge². This has introduced 30-day readmission as a quality metric to determine the quality of care provided to patients by hospitals. The law suggests that the re-imbursements to individual hospitals should be determined by a factor, calculated by hospital’s expected versus observed 30-day re-admission rates. Agency for Healthcare Research and Quality has started a project, named Project Re-Engineered Discharge (RED), which focuses on ensuring patient education at the time of hospital discharge in order to reduce hospital readmissions¹⁷. In the era of diminishing reimbursements, it is easily predictable that in near future, hospitals may be penalized for early readmissions after discharge. Recently, there has been increased focus on reducing all hospital readmissions, including post-surgical patients^18,19.

Readmission for surgical patients is fundamentally different from medical patients. While the most common reason for readmission for majority of the medical patients is the primary diagnosis itself¹⁶, the majority (70.5%) of surgical readmissions are due to post-operative complications³. The three most common procedures associated with readmissions include leg amputations, colorectal resections and peripheral vascular bypasses³. Surgical patients have the same underlying medical comorbidities as medical patients, but in addition, they undergo surgical operations which themselves are associated higher risk of readmissions. Since majority of the surgical procedures are planned, it implies that there is an opportunity to prevent complications and readmissions.

Patients with peripheral arterial disease (PAD) represent a patient population with significant medical comorbidities. With recent advances in the field of endovascular surgery, majority of the patients with PAD are treated with endovascular interventions. Currently, patients who require peripheral arterial bypasses are generally a select group of patients who have the most severe form of disease, which is not amenable to endovascular interventions. Our study shows that the risk of re-admission correlates with the severity of PAD: incidence of readmission was the highest among the patients who presented with the most severe forms of PAD i.e. tissue loss (19.8%) and rest pain (16.9%) when compared to patients with the early states of disease (claudication, 10.5%). These findings correlate with the fact that patients with high-risk physiologic factors were more likely to get readmitted (20%) when compared to those with absence of these factors (14.9%). Similarly, patients with diabetes, congestive heart failure, hypertension and open wounds at the time of operation were more likely to be re-admitted. Intra-operative factors associated with increased risk of readmission included longer operative times and high American Society of Anesthesiology (ASA) scores. While most of these variables represent patients with significant systemic disease and are likely not preventable, it may represent a target population for maximizing optimal treatment of these co-morbidities before the operation. It is important to mention that some patients who undergo lower extremity bypass procedures may be readmitted for planned staged operations such as foot procedures or for treatment of contralateral limb ischemia. Our study includes only those readmissions, which were not planned.

A surgical operation is a physiologic stress for the body. Our data shows that intra-operative factors associated with increased risk of readmission include longer operative times and higher ASA scores. Both of these factors suggest that patients who need readmissions represent a subgroup of patients who are generally sicker than other patients and probably have the most severe form of atherosclerotic disease, necessitating longer operative times. Our study identifies several post-operative factors, which put a patient at a higher risk for re-admission. Unlike heart failure patients¹, risk of readmission was slightly higher for patients who had longer lengths of hospital stay. Incidence of post-operative bleeding, myocardial infarction, stroke, need for reoperation, amputation, pneumonia, wound disruption, superficial wound infection, deep wound infection, urinary tract infections and progressive renal insufficiency were all associated with increased risk of 30-day readmission (Table I). The majority of these factors are preventable. Other authors have identified similar risk factors for hospital readmission after general surgery operations²⁰. Unfortunately, NSQIP database did not capture the reason for admission for majority of these patients (Table II). Superficial wound infections and wound related complications seem to be the most common reason for readmission among the patients, in whom the reason for readmission was documented. This is contrary to other studies, looking at readmission rates for general surgery operations e.g. a recent analysis of patients undergoing bowel surgery showed that the post-operative infections were responsible for only 6.4% of early readmissions¹⁶. However, our findings are similar to the studies, looking at readmission rates after coronary artery bypass graft surgery²¹ which identify wound infections to be the most common cause of hospital readmissions with similar incidence (16.5%) of early readmission. In addition, our study clearly shows that patients with high-risk physiologic factors are more likely to get readmitted within 30 days of the index operation, highlighting the importance of recognizing the importance of optimal management of systemic diseases in patients with limb ischemia. It is possible that patients with severe coronary artery heart disease and lower extremity arterial disease represent a cohort of patients with the most severe form of atherosclerosis and are at a higher risk for developing wound infections.

In addition to financial loss for the hospitals, there are several other factors, which reflect the negative impact of re-admissions. An early re-admission after discharge from hospital affects patients’ quality of life and delays recovery and return to work. It also puts constraints on hospitals’ resources. Resources consumed for care of readmitted patients can be used to treat other patients.

Although return to operating room after the index operation of lower extremity revascularization has been previously shown to be a predictor for early readmission in institutional studies²², our data is based on patients from multiple institutions across the country and hence represents a broader picture. Our analysis clearly shows that the patients who needed to be brought back to the operating room were at a 5.95-fold higher risk of 30-day readmission. This is an important finding, as potentially all unplanned returns to the operating room are preventable and represent an area for improvement. Unfortunately, NSQIP database lacks information on significant findings during re-operations. It is a well-established fact that most of the unplanned returns to the operating rooms for general surgery patients are contributed to technical failures related to hemostasis, wound closure and anastomosis^{9, 23} It is difficult to establish when an unplanned return to the operating room is truly preventable. It is fair to state that any such event is likely multifactorial, and errors in clinical judgment and or technical failures have at least some role to play in such events From quality standpoint, any unplanned return to the operating room is easily identifiable in administrative and clinical databases and hence, it can be an attractive quality matrix. However, there are several downsides to using re-operation as a surrogate for quality. First, re-operations can be easily affected by different case-mix index of different institutions. Second, thresholds for re-operation may be variable among different surgeons. For example, some surgeons may chose to manage a local wound hematoma conservatively and others may choose aggressive approach to drain the incisional hematoma, which may reflect badly when reoperation is used as a quality index, but may in fact help the patient by releasing the pressure from hematoma and encouraging early ambulation, decrease length of hospital stay and lower the need for pain medications.

This study has several limitations. It is a retrospective analysis of a large, national surgical database. Analyzed variables are limited to those, which are recorded by NSQIP. Current NSQIP dataset does not capture the findings encountered during re-operation. These findings are restricted to only those hospitals, which are participating in NSQIP. Lastly, this database does not record outcomes beyond thirty days, and any patient who need to re-operations and or readmissions beyond this time period will not be captured in analyses. The strength of this study is that NSQIP database is the largest and the most comprehensive surgical database, available to surgeons across the country. The American College of Surgeons ensures quality of this dataset and numerous studies based on this database have produced reproducible results.

We summarize by stating that re-admission to hospital after a surgical operation can be considered an indicator of poor quality health care Our analysis reveals several factors that are associated with increased readmission rates after lower extremity revascularization procedures. The aim of this study is to focus on the factors that are generally deemed preventable. Unplanned reoperation after surgery is considered potentially preventable and is associated with a 5.95-fold risk for re-admission. Peripheral arterial bypasses require attention to detail and technical finesse. Every effort should be made to minimize technical errors during the operation. Any decision to bring the patient back to the operating room should be made after a well-thought process, using sound clinical judgment.

Appendix: Definitions (from the User Guide for the 2013 ACS-NSQIP Participant Use Data File and the User Guide 2013 for the ACS-NSQIP Procedure Targeted Participant Used Data File)

I. High Risk Criteria

a) High Risk Factors, Physiologic: Select “Yes” if one or more of the following apply to the case:

1. End Stage Renal Disease (ESRD)

   Examples of ESRD include:

   • Patient is documented to have stage 5 chronic kidney disease
   • Patient is documented to have end-stage renal disease (ESRD)
   • Patient’s glomerular filtration rate (GFR) is less than 15 mL/min per 1.73 m2.
   • Patient is documented to be on chronic renal replacement therapy (hemodialysis or peritoneal dialysis)

2. Age greater than 80

3. New York Heart Association (NYHA) congestive heart failure (CHF) Class III/IV: For patients with a pre-operative diagnosis of congestive heart failure within 30 days prior to surgery please refer to the pre-operative note for the New York Heart Association functional class. If the patient has CHF and the NYHA functional class is not documented, please refer to the pre-operative notes for the severity of the patient’s symptoms to determine the NYHA functional class. Symptoms of CHF include: cough, dyspnea, chest pain/angina, peripheral edema.

   • Class I – symptoms of HF only at activity levels that would limit normal individuals
   • Class II – symptoms of HF with ordinary exertion
   • Class III – symptoms of HF with less than ordinary exertion
   • Class IV – symptoms of HF at rest

4. Left Ventricular Ejection Fraction (LVEF) < 30%: Documentation in the patient’s medical record of left ventricular ejection fraction less than 30%.

   Data sources include any of the following:

   • Echocardiogram
   • Coronary angiogram
   • Radionucleotide ventriculography (radionucleotide angiography or RNA, radionucleotide cine angiography or RNCA)
   • Multiple gated cardiac blood pool imaging (MUGA
   • Equilibrium radionucleotide angiography (ERNA)
   • ECG gated single photon emission tomography (SPECT)
   • Cardiovascular MRI (CMRI)
   • Coronary CT angiography

5. Unstable angina: Documentation in the patient’s medical record of a history of unstable angina (or “anginal equivalent”) within 30 days prior to surgery.

   Symptoms include at least 1 of the following:

   • Angina at rest, lasting >20 minutes
   • New onset angina that markedly limits physical activity
   • Increasing angina that is more frequent, longer in duration, or occurs with less exertion than previous angina

6. Recent myocardial infarction (MI) (within 30 days): History of myocardial infarct (ST-elevation or non-ST-elevation MI) in the 30 days prior to the surgery documented in the patient’s medical record.

b) High Risk Factors, Anatomic: Select the anatomic factor for surgery:(1) Prior ipsilateral bypass involving currently treated segment: Surgeon’s pre-operative note indicates the patient previously underwent a bypass procedure on the same arteries as the current procedure (i.e. axillary-femoral, aorto-iliac, aorto-femoral, femoral-femoral, femoral-popliteal, femoral-tibial, femoral-pedal arterial bypass).

(2) Prior ipsilateral percutaneous intervention involving currently treated segment: History of a previous percutaneous angioplasty or stent treatment of the same lower extremity artery as the current procedure is planned to treat.

(3) None/Not documented

II. Significant Post Operative Bleeding:

Any transfusion (including autologous) of packed red blood cells or whole blood given from the time the patient leaves the operating room up to and including 72 hours post operatively. Bleeding transfusion entered for five or more units of packed red blood cell units in the postoperative period including hanging blood from the operating room that is finished outside of the OR. If the patient received shed blood, autologous blood, cell saver blood or pleurovac postoperatively, this is counted if greater than four units. The blood may be given for any reason.

III. Postoperative Myocardial Infarction:

A new transmural acute myocardial infarction occurring during surgery or within 30 days as manifested by new Q–waves on EKG.

IV. Postoperative Stroke:

Patient develops an embolic, thrombotic, or hemorrhagic vascular accident or stroke with motor, sensory or cognitive dysfunction (e.g. hemiplegia, hemiparesis, aphasia, sensory deficit, impaired memory) that persists for 24 or more hours. If a specific time frame for the dysfunction is not documented in the medical record, but there is a diagnosis of stroke, assign the occurrence, unless documentation specifically states that the motor, sensory or cognitive dysfunction resolved.

V. Deep Wound infection:

Deep incisional surgical site infection that occurs within 30 days after the operation and the infection appears to be related to the operation and infection involved deep soft tissues (e.g. fascial and muscle layers) of incision and at least one of the following:

• Purulent drainage from the deep incision but not from the organ/space component of the surgical site

• A deep incision spontaneously dehices or is deliberately opened by a surgeon while the patient has at least one of the following signs or symptoms:

  • Fever (>38 C)
  • Localized Pain
  • Tenderness, unless the site is culture-negative
  • An abscess or other evidence of infection involving the deep incision is found on direct examination, during reoperation, or by histopathologic or radiologic examination
  • Diagnosis of a deep surgical site infection by a surgeon or attending physician.

Note: An organ or space surgical site infection that drains through the incision is reported as a deep incisional surgical site infection.

VI. Loss of bypass patency:

Image proven graft thrombosis or clinically evident graft thrombosis.

VII. Readmission:

Any unplanned readmission (to the same or another hospital) for a post-operative occurrence likely related to the principal surgical procedure within 30 days of the procedure.