Surgical Quality Improvement Program Dataset Research Articles

Trainee-associated outcomes in laparoscopic colectomy for cancer: propensity score analysis accounting for operative time, procedure complexity and patient comorbidity.

Surgical trainee association with operative outcomes is controversial. Studies are conflicting, possibly due to insufficient control of confounding variables such as operative time, case complexity, and heterogeneous patient populations. As operative complications worsen long-term outcomes in oncologic patients, understanding effect of trainee involvement during laparoscopic colectomy for cancer is of utmost importance. Here, we hypothesized that resident involvement was associated with worsened 30-day mortality and 30-day overall morbidity in this patient population. Patients undergoing laparoscopic colectomy for oncologic diagnosis from 2005 to 2012 were assessed using the American College of Surgeons National Surgical Quality Improvement Program dataset. Propensity score matching accounted for demographics, comorbidities, case complexity, and operative time. Attending only cases were compared to junior, middle, chief resident, and fellow level cohorts to assess primary outcomes of 30-day mortality and 30-day overall morbidity. A total of 13,211 patients met inclusion criteria, with 4075 (30.8%) cases lacking trainee involvement and 9136 (69.2%) involving a trainee. Following propensity matching, junior (PGY 1-2) and middle level (PGY 3-4) resident involvement was not associated with worsened outcomes. Chief (PGY 5) resident involvement was associated with worsened 30-day overall morbidity (15.5 vs. 18.6%, p=0.01). Fellow (PGY>5) involvement was associated with worsened 30-day overall morbidity (16.0 vs. 21.0%, p<0.001), serious morbidity (9.3 vs. 13.5%, p<0.001), minor morbidity (9.8 vs. 13.1%, p=0.002), and surgical site infection (7.9 vs. 10.5%, p=0.006). No differences were seen in 30-day mortality for any resident level. Following propensity-matched analysis of cancer patients undergoing laparoscopic colectomy, chief residents, and fellows were associated with worsened operative outcomes compared to attending along cases, while junior and mid-level resident outcomes were no different. Further study is necessary to determine what effect the PGY surgical trainee level has on post-operative morbidity in cancer patients undergoing laparoscopic colectomy in the context of multiple collinear factors.

Validating the Thoracic Revised Cardiac Risk Index Following Lung Resection

The Thoracic Revised Cardiac Index (ThRCRI) is a tool that differentiates patients who may proceed to lung resection (classes A or B) from those who should receive additional cardiac evaluation (classes C or D). This study aims to describe the ability of the ThRCRI to stratify patients based on major cardiac complication rates using a large multi-institutional dataset. Patients undergoing lobectomy or pneumonectomy were identified in the American College of Surgeons National Surgical Quality Improvement Program dataset from 2005 to 2012. Patients were grouped into 4 risk classes based on a summary score of preoperative risk factors: ischemic heart disease, cerebrovascular disease, renal comorbidity, and pneumonectomy. The primary outcome was the incidence of perioperative major cardiac complication in each of the 4 risk classes. Of the 4,625 patients identified, the majority underwent surgery for malignant disease (78%) and had an open procedure (70%). Among ThRCRI risk factors, 9% of patients had ischemic heart disease, 7% had cerebrovascular disease, 2% had renal comorbidity, and 6% underwent pneumonectomy. Incidence of cardiac complication in all patients was 2%. Incidence of cardiac complication within risk classes A, B, C, and D were 1%, 3%, 9%, and 4%, respectively (p < 0.01). Using a large multi-institutional dataset, the ThRCRI can differentiate patients at higher risk for cardiac complication following lung resection (classes C and D) and can be a useful preoperative instrument. The ThRCRI may allow for identifying patients who would benefit from additional cardiac evaluation.

An Analysis of Independent Variables Affecting Surgical Outcomes in Patients Undergoing Repair of Maxillofacial Trauma: An American College of Surgeons National Surgical Quality Improvement Program Study.

Facial fractures, from straightforward closed nasal reductions to complex pan facial fractures, are commonly encountered in the Plastic Surgical community. However, very little has been discussed in the literature regarding the outcomes of facial fractures relating to contributing factors. Our aim was to evaluate a battery of independent variables in order to identify, which, if any, factors correlate with suboptimal outcomes in patients who have undergone facial fracture surgery. Under the data use agreement of the American College of Surgeons public use files of the NSQIP, patients involving repair of facial fractures, Current Procedural Terminology codes 21310 to 21470 inclusive, were queried. The outcomes examined included: wound dehiscence, superficial surgical site infection, deep surgical site infection, readmission, open wound/wound infection and return to the operating room. There were 2069 facial fracture patients in the National Surgical Quality Improvement Program datasets (2005-2013). Thirteen perioperative risk factors and patient comorbidities were evaluated for correlation with the 6 outcomes. Of the 6 outcomes evaluated, open wound/wound infection was the most prevalent outcome (6%). Factors statistically significant for presence of open wound/wound infection were American Society of Anesthesiology classification (P = 0.002), presence of bleeding disorder (P = 0.008), emergency patient (P = 0.001), chronic alcohol use (P = 0.002), and chronic steroid use (P = 0.034). Several factors correlated with presence of an open wound/wound infection; however, variables such as diabetes and active tobacco use, which are often thought to contribute to wound infections, were shown to be statistically nonsignificant. Although this study was limited by its observational nature, these data may indicate a change in perception of the factors correlated with wound infections.

Incidence and Risk Factors of Postoperative Hematoma Requiring Reoperation in Single-level Lumbar Fusion Surgery.

Retrospective cohort. The purpose of the present study was to examine the incidence and risk factors for postoperative hematoma requiring reoperation in patients undergoing single-level lumbar fusion surgery. Postoperative hematoma can cause devastating neurological consequences after spine surgery. Risk factors for hematoma in specific spine procedures have not been well established. A cohort of patients undergoing single-level lumbar fusion surgery was constructed from the 2012 to 2013 American College of Surgeons National Surgical Quality Improvement Program dataset using Current Procedural Terminology codes (22533, 22558, 22612, 22630, and 22633). In cases requiring reoperation within 30 days after initial surgery, postoperative hematoma was identified using the ICD-9 code 998.1. Risk factors for postoperative hematoma were assessed with logistic regression modeling. Of 5280 patients undergoing single-level lumbar fusion surgery, 27 patients (0.5%) developed a postoperative hematoma requiring reoperation for hematoma evacuation. A heightened incidence of postoperative hematoma was found in patients who were smokers (1.0% vs. 0.4% for nonsmokers, P = 0.016) or who had a diagnosis of bleeding disorder (3.8% vs. 0.5% for those without bleeding disorder, P = 0.007). Multivariate logistic regression analysis indicated that the adjusted odds ratios for postoperative hematoma associated with smoking and bleeding disorder were 3.34 (95% confidence interval, 1.15-9.71) and 10.2 (95% confidence interval, 1.9-54.8), respectively. Smoking and bleeding disorder appear to be major risk factors for postoperative hematoma requiring reoperation after single-level lumbar fusion surgery. Intervention programs targeting patients with these risk factors are needed to reduce their excess risk of postoperative hematoma. 3.

Does Receiving a Blood Transfusion Predict for Length of Stay in Children Undergoing Cranial Vault Remodeling for Craniosynostosis? Outcomes Using the Pediatric National Surgical Quality Improvement Program Dataset

Recent interventions have aimed at reducing the need for blood transfusions in the perioperative period in patients with craniosynostosis undergoing cranial vault remodeling. However, little is known regarding whether the receipt of a blood transfusion influences the length of hospital stay. The purpose of this study was to assess whether the receipt of a blood transfusion in patients undergoing cranial vault remodeling is associated with an increased length of stay. To address the research purposes, we designed a retrospective cohort study using the 2014 Pediatric National Surgical Quality Improvement Program (NSQIP Peds) dataset. The primary predictor variable was whether patients received a blood transfusion during cranial vault remodeling. The primary outcome variable was length of hospital stay after the operation. The association between the receipt of blood transfusions and length of stay was assessed using the Student t test. The association between other covariates and the outcome variable was assessed using linear regression, analysis of variance, and the Tukey test for post hoc pair-wise comparisons. The sample was composed of 756 patients who underwent cranial vault remodeling: 503 who received blood transfusions and 253 who did not. The primary predictor variable of blood transfusion was associated with an increased length of stay (4.1days vs 3.0days, P= .03). Other covariates associated with an increased length of stay included race, American Society of Anesthesiologists status, premature birth, presence of a congenital malformation, and number of sutures involved in craniosynostosis. The receipt of a blood transfusion in the perioperative period in patients with craniosynostosis undergoing cranial vault remodeling was associated with an increased length of stay.

Minimally invasive hepatectomy conversions: An analysis of outcomes.

430 Background: Surgical approach has been described as a factor influencing morbidity in patients undergoing hepatectomy. Minimally invasive surgery (MIS) converted to open has previously been included in an open hepatectomy cohort of the National Surgical Quality Improvement Program (NSQIP) dataset for analysis. This study sought to compare MIS conversion to open hepatectomy cases and analyze factors leading to conversion. Methods: The NSQIP data set was queried for patients undergoing hepatectomy for liver disease between January 1, 2014 and December 31, 2014. Cases were divided into three cohorts: completely MIS (laparoscopic or robotic), open, or planned MIS converted to open approach according to the dataset definitions. Demographic, clinical-pathologic, and perioperative variables were described and short-term postoperative outcomes were compared between groups. Results: A total of 2,884 patients were identified. Open approach was performed most commonly (80.96%, n = 2,335), followed by minimally invasive (15.05%, n = 434) and MIS converted to open (3.99%, n = 115). Patients undergoing conversion were more likely to have higher BMI, hypertension, and hepatocellular carcinoma. Conversion patients suffered from significantly higher morbidity (p &lt; 0.001) compared to MIS, but had comparable outcomes to open procedures. Conclusions: MIS hepatectomy has been shown to have lower morbidity but similar mortality to open hepatectomy. Conversion procedures appear to have similar outcomes to open procedures, with greater morbidity compared to MIS. Differences in morbidity from previous studies may be partially attributed to the inclusion of conversions in the open hepatectomy cohort.

Determining the Optimal Quantitative Threshold for Preoperative Albumin Level Before Elective Colorectal Surgery

BackgroundHypoalbuminemia is associated with adverse surgical outcomes. A minimum threshold and the impact of incrementally decreasing albumin remain undefined for colorectal surgery patients. Study DesignThe 2011–2013 National Surgical Quality Improvement Program (NSQIP) dataset was queried for patients undergoing elective colorectal surgery. Multivariable regression analyses with restricted cubic splines (RCS) were used to examine the adjusted association between preoperative serum albumin level and the incidence of complications and to establish an optimal threshold. RCS allows for flexible evaluation in multivariable models without having to assume a specific relationship a priori. ResultsSixteen thousand one hundred forty-five patients met study criteria. RCS analysis demonstrated an inflection point at serum albumin level of 3.9 mg/dL. Patients with preoperative albumin <3.9 mg/dL vs. albumin ≥3.9 mg/dL had a higher likelihood of experiencing a major complication (odds ratio (OR) = 1.18, confidence interval (CI) 1.07–1.30, p = 0.0007) or any complications (OR 1.18, CI 1.08–1.29, p = 0.0002,) and had a lengthened hospital stay (p < 0.001). ConclusionsThis study objectively determines that a threshold preoperative serum albumin of ≥3.9 mg/dL is associated with improved outcomes in elective colorectal surgery patients. Each 0.5 mg/dL decrease in albumin was progressively associated with increased risk for complications. Identifying a minimum albumin threshold has implications in perioperative optimization of patients undergoing colorectal surgery.

Who Should Get Extended Thromboprophylaxis After Bariatric Surgery?: A Risk Assessment Tool to Guide Indications for Post-discharge Pharmacoprophylaxis.

To determine the risk factors for 30-day postdischarge venous thromboembolism (VTE) after bariatric surgery and to identify potential indications for extended pharmacoprophylaxis. VTE is among most common causes of death after bariatric surgery. Most VTEs occur after hospital stay; still a few patients receive extended pharmacoprophylaxis postdischarge. From American College of Surgeons-National Surgical Quality Improvement Program, we identified 91,963 patients, who underwent elective primary and revisional bariatric surgery between 2007 and 2012. Regression-based techniques were used to create a risk assessment tool to predict risk of postdischarge VTE. The model was validated using the 2013 American College of Surgeons-National Surgical Quality Improvement Program dataset (N = 20,575). Significant risk factors were used to create a user-friendly online risk calculator. The overall 30-day incidence of postdischarge VTE was 0.29% (N = 269). In those experiencing a postdischarge VTE, mortality increased about 28-fold (2.60% vs 0.09%; P < 0.001). Among 45 examined variables, the final risk-assessment model contained 10 categorical variables including congestive heart failure, paraplegia, reoperation, dyspnea at rest, nongastric band surgery, age ≥60 years, male sex, BMI ≥50 kg/m, postoperative hospital stay ≥3 days, and operative time ≥3 hours. The model demonstrated good calibration (Hosmer-Lemeshow goodness-of-fit test, P = 0.71) and discrimination (c-statistic = 0.74). Nearly 2.5% of patients had a predicted postdischarge VTE risk >1%. More than 80% of post-bariatric surgery VTE events occurred post-discharge. Congestive heart failure, paraplegia, dyspnea at rest, and reoperation are associated with the highest risk of post-discharge VTE. Routine post-discharge pharmacoprophylaxis can be considered for high-risk patients (ie, VTE risk >0.4%).

Postoperative Complications and Readmission Rates Following Surgery for Cerebellopontine Angle Schwannomas.

To investigate the 30-day postoperative complication, readmission, and reoperation rates following surgery for cerebellopontine angle (CPA) schwannomas. Cross-sectional analysis. National surgical quality improvement program dataset (NSQIP) 2009 through 2013. All surgical cases with an International Classification of Diseases, 9th edition, Clinical Modification (ICD-9-CM) diagnosis code of 225.1, benign neoplasms of cranial nerves, and one of the following current procedural terminology (CPT) codes, were included: 61616, 61526, 61530, and 61520. Surgical resection as indicated by the CPT codes above. Demographics, comorbidities, 30-day postoperative complications, readmission rate, and reoperation rate. Overall, 404 cases were identified, of which 42.6% were men. The average age was 51 years. Comorbidities were present in 45.3%. NSQIP-tracked complications occurred in 9.7% of patients. Most common complications were wound infections including surgical-site infection and wound dehiscence (11 patients, 2.7%), sepsis (10 patients, 2.5%), blood loss (nine patients, 2.2%), and deep vein thrombosis (DVT; seven patients, 1.7%). Mortality occurred in four patients (1.0%). The complication rate was statistically higher in patients with comorbidities versus those without (10.2% versus 4.1%, p = 0.04). Patients with complications were more likely to undergo reoperation (2.5% with versus 4.1% without, p = 0.001). Unplanned readmissions occurred in 41 cases (10.1%) and reoperations occurred in 23 patients (5.7%). Most common NSQIP-tracked complications in excision of CPA neoplasms are infections, sepsis, blood loss, and deep vein thrombosis (DVT). Further, investigation of patients with unplanned readmission and reoperation are warranted. Neurotologists need to take an active role in the data to be gathered in the NSQIP database as it relates to vestibular schwannomas.

Predictors of 30-day perioperative morbidity and mortality of unruptured intracranial aneurysm surgery

IntroductionLarge-scale studies examining the incidence and predictors of perioperative complications after surgical clipping of unruptured intracranial aneurysms (UIA) using nationally representative prospectively collected data are lacking in the literature. MethodsUsing the American College of Surgeons’ National Surgical Quality Improvement Program (ACS-NSQIP) dataset, we conducted a retrospective analysis of the complications experienced by patients that underwent surgical management of a UIA between the years of 2007 and 2013. The primary outcomes of interest were mortality within the 30-day perioperative period and adverse discharge disposition to a location other than home. Predictors of morbidity and mortality were elucidated using multivariable logistic regression analyses controlling for available patient demographic, comorbidity, and operative characteristics. Results662 patients were identified in the ACS-NSQIP dataset for operative management of an unruptured aneurysm. The observed rates of 30-day mortality and adverse discharge disposition were 2.27% and 19.47%, respectively. A hundred and eight (16.31%) patients developed at least one major complication. On multivariable analysis, death within 30days was significantly associated with increased operative time (OR 1.005 per minute, 95% CI 1.002–1.008) and chronic preoperative corticosteroid use (OR 28.4, 95% CI 1.68–480.42), whereas major complication development was associated with increased operative time (OR 1.004 per minute, 95% CI 1.002–1.006), age (OR 1.017 per year, 95% CI 1–1.034), preoperative dependency (OR 3.3, 95% CI 1.16–9.40) and diabetes mellitus (OR 2.89, 95% CI 1.45–5.75). Lastly, increasing age (OR 1.017 per year, 95% CI 1–1.034) as well as ASA Class 3 (OR 1.73, 95% CI 1.08–2.77) and 4 (OR 2.28, 95% CI 1.1–4.72) were independent predictors of discharge to a location other than home. ConclusionOur study yields morbidity and mortality benchmarks for UIA surgery in a representative, national surgical registry. It will hopefully aid in recognizing those patients at greater risk for postoperative complications following surgical management, leading to appropriate changes in treatment strategies for this selected group of patients.

Primary Anastomosis versus Ostomy after Colon Resection during Debulking of Ovarian Carcinomatosis: A NSQIP Analysis

Ovarian carcinomatosis poses a dilemma for the surgeon. When resecting colon for tumor invasion, one must decide between diversion and primary anastomosis (PA). We examined the National Surgical Quality Improvement Program to determine whether PA associated with more complications than ostomy. The National Surgical Quality Improvement Program dataset was queried for patients with ovarian carcinomatosis between 2007 and 2012. Current Procedural Terminology codes were used to further identify patients undergoing colectomy with PA or ostomy. Logistic regression was used to evaluate 30-day morbidity and mortality. The 1013 ovarian carcinomatosis patients who underwent elective colon surgery were divided into primary repair (n = 453, 43.5%) or ostomy (n = 586, 56.5%) groups. Preoperative demographics were similar; however, ostomy patients had more severe preoperative laboratory derangements. The 30-day mortality and postoperative transfusion requirements were higher in the ostomy group. On multivariate analysis controlling for confounders, the differences were no longer significant. In conclusion, 30-day mortality and postoperative complications were increased in the ostomy group. Given the laboratory derangements in this group, this may reflect tendency to allocate ostomies to more ill patients. Primary repair in a selected population does not worsen outcomes. Prospective evaluation would help determine the impact of PA in the ovarian carcinomatosis population.

Effect of body mass index percentile on pediatric gastrointestinal surgery outcomes

Background/purposePediatric obesity is an important public health concern, yet its effect on surgical outcomes is poorly understood. The purpose of this study was to determine if age and gender-specific body mass index (BMI) percentile influences complications and hospital resource utilization following pediatric gastrointestinal surgeries. MethodsPatients aged ≥2 to <18years who underwent appendectomy or other gastrointestinal operations were identified in the 2012–2013 Pediatric National Surgical Quality Improvement Program datasets. Age- and gender-specific pediatric BMI percentiles were calculated. Patients who underwent appendectomy (n=9606) and those undergoing all other intestinal operations (n=2664) were evaluated as separate cohorts. ResultsIn the appendectomy cohort, frequency of any complication increased with BMI category (normal weight 4.5%, overweight 5.3%, obese 5.7%, morbidly obese 7.3%, overall 5.0%, p=0.014). In multivariate analysis, there was a quadratic association between BMI percentile and increased frequency of superficial incisional infection, unplanned tracheal intubation, and longer operative duration. In the intestinal surgery cohort, BMI percentile was not a predictor of any individual complication or any measure of hospital utilization. ConclusionsAge- and gender-specific BMI percentile was associated with increased risk of complications and longer operative duration in patients undergoing appendectomy but not other intestinal operations.

Wound classification in pediatric surgical procedures: Measured and found wanting

PurposeSurgical wound classification has emerged as a measure of surgical quality of care, but scant data exist in the era of minimally invasive procedures, especially in children. The aim of this study is to examine the surgical site infection (SSI) rate by wound classification during common pediatric surgical procedures. MethodsA retrospective analysis of the 2013 Pediatric–National Surgical Quality Improvement Program (Peds-NSQIP) dataset was conducted. Patients undergoing pyloromyotomy, cholecystectomy, ostomy reversal, and appendectomy were included. Wound classification, SSI rate, reoperation, and readmission were analyzed. ResultsA total of 10,424 records were included. Pyloromyotomy, a clean case, had a 0.7% SSI rate, while ostomy reversal, a clean contaminated case, had an SSI in 6.9% of cases. Appendectomy for nonperforated acute appendicitis and laparoscopic cholecystectomy for cholecystitis, both contaminated cases, had SSI rates of 2.1% and <1%, respectively. Appendectomy for perforated appendicitis, a dirty procedure, had a 9.1% SSI rate, below the expected >40% for dirty cases. Reoperations and readmission rates ranged from <1% to 9% and increased with case complexity. ConclusionCurrent wound classifications systems do not reflect surgical risk in children and remain questionable tools for benchmarking surgical care in children. Role of readmissions and reoperations as quality of care indices needs further investigation.

Impact of frailty on outcomes in geriatric femoral neck fracture management: An analysis of national surgical quality improvement program dataset

IntroductionFrailty is a clinical state of increased vulnerability resulting from aging-associated decline in physiologic reserve. Hip fractures are serious fall injuries that affect our aging population. We retrospectively sought to study the effect of frailty on postoperative outcomes after Total Hip Arthroplasty (THA) and Hemiarthroplasty (HA) for femoral neck fracture in a national data set. MethodsNational Surgical Quality Improvement Project dataset (NSQIP) was queried to identify THA and HA for a primary diagnosis femoral neck fracture using ICD-9 codes. Frailty was assessed using the modified frailty index (mFI) derived from the Canadian Study of Health and Aging. The primary outcome was 30-day mortality and secondary outcomes were 30-day morbidity and failure to rescue (FTR). We used multivariate logistic regression to estimate odds ratio for outcomes while controlling for confounders. ResultsOf 3121 patients, mean age of patients was 77.34 ± 9.8 years. The overall 30-day mortality was 6.4% (3.2%-THA and 7.2%-HA). One or more severe complications (Clavien-Dindo class-IV) occurred in 7.1% patients (6.7%-THA vs.7.2%-HA). Adjusted odds ratios (ORs) for mortality in the group with the higher than median frailty score were 2 (95%CI, 1.4–3.7) after HA and 3.9 (95%CI, 1.3–11.1) after THA. Similarly, in separate multivariate analysis for Clavien-Dindo Class-IV complications and failure to rescue 1.6 times (CI95% 1.15–2.25) and 2.1 times (CI95% 1.12–3.93) higher odds were noted in above median frailty group. ConclusionsmFI is an independent predictor of mortality among patients undergoing HA and THA for femoral neck fracture beyond traditional risk factors such as age, ASA class, and other comorbidities. Levels of evidenceLevel II.

Is expedited early discharge following elective surgery for colorectal cancer safe? An analysis of short-term outcomes.

Enhanced recovery pathways have become standard practice after elective colorectal surgery to improve postoperative care while reducing length of stay in hospital. However, there is concern that early discharge may result in increased rates of adverse events including readmission. This study aims to determine whether it is safe to discharge patients on postoperative day 1 or 2 if they have undergone an elective colorectal operation for cancer. The 2012 American College of Surgeons National Surgical Quality Improvement Program dataset was used. The study included patients who underwent elective colorectal cancer surgery and were discharged on postoperative day (POD) 1 or 2 (expedited early discharge) versus POD 3 or 4 (standard early discharge). Patients who had metastases, concurrent procedures including ostomy creation, or died during admission were excluded. Primary outcomes were 30-day adverse events (serious complications, mortality and reoperations) and readmission rates, which were analyzed using multivariable regression. A total of 305 and 2277 patients were identified in the expedited and standard early discharge groups, respectively. There were 6 (1.97%) adverse events and 16 (5.56%) readmissions in the expedited group, compared to 59 (2.59%) and 135 (6.24%) in the standard group. No statistical difference was found between the cohorts with respect to 30-day adverse events (OR 0.93, p=0.87, 95% CI [0.41-2.12]) or readmission rate (OR 1.03, p=0.90, 95% CI [0.61-1.76]). Patients discharged by POD 2 after elective oncologic colon resections did not have significantly more adverse events or readmissions compared to patients discharged later. Select patients may be safely discharged earlier.

Predictors of 30-day readmission after mastectomy: A multi-institutional analysis of 21,271 patients.

Recent healthcare legislation has made unplanned hospital readmission an important metric of health care quality, and current efforts center on reducing this complication in order to avoid fiduciary penalties. There is currently a paucity of data delineating risk factors for readmission following mastectomy. To this end, we sought to develop a predictive model of unplanned readmissions following mastectomy. The 2011 and 2012 National Surgical Quality Improvement Program (NSQIP) datasets were retrospectively queried to identify patients who underwent mastectomy. Multivariate logistic regression modeling was used to identify risk factors for readmission. Of 21,271 patients meeting inclusion criteria, 1,190 (5.59%) were readmitted. The most commonly cited reasons for readmission included surgical site complications (32.85%), infection not localized to the surgical site (2.72%), and venous thromboembolism (4.39%). Independent predictors of readmission included BMI, active smoking status, and skin-sparing mastectomy. Significantly, concurrent breast reconstruction and bilateral mastectomy were not independent predictors of readmission. This is the first study of readmission rates after mastectomy. Awareness of specific risk factors for readmission, particularly those that are modifiable, may serve to identify and manage high risk patients, aid in the development of pre- and postoperative clinical care guidelines, and ultimately improve patient care.

A139 - Who Should Get Extended Thromboprophylaxis After Bariatric Surgery? A Risk Assessment Tool to Guide Indications for Post-Discharge Pharmacoprophylaxis

To determine the risk factors for 30-day postdischarge venous thromboembolism (VTE) after bariatric surgery and to identify potential indications for extended pharmacoprophylaxis. VTE is among most common causes of death after bariatric surgery. Most VTEs occur after hospital stay; still a few patients receive extended pharmacoprophylaxis postdischarge. From American College of Surgeons-National Surgical Quality Improvement Program, we identified 91,963 patients, who underwent elective primary and revisional bariatric surgery between 2007 and 2012. Regression-based techniques were used to create a risk assessment tool to predict risk of postdischarge VTE. The model was validated using the 2013 American College of Surgeons-National Surgical Quality Improvement Program dataset (N = 20,575). Significant risk factors were used to create a user-friendly online risk calculator. The overall 30-day incidence of postdischarge VTE was 0.29% (N = 269). In those experiencing a postdischarge VTE, mortality increased about 28-fold (2.60% vs 0.09%; P 1%. More than 80% of post-bariatric surgery VTE events occurred post-discharge. Congestive heart failure, paraplegia, dyspnea at rest, and reoperation are associated with the highest risk of post-discharge VTE. Routine post-discharge pharmacoprophylaxis can be considered for high-risk patients (ie, VTE risk >0.4%).

A127 - Impact of Metabolic Surgery for Patients With Bmi Less Than 35 Kg/M2 on Health-Related and Eating Quality of Life

fall into this low BMI category. Methods: Data were retrieved from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) dataset (2005-2013). Inclusion criteria included: adult patients aged Z18 years, BMI Z25 but o35 kg/m, who were being treated with medications for T2D and underwent elective primary bariatric surgical procedures including Roux-en-Y gastric bypass (RYGB), adjustable gastric banding (AGB), sleeve gastrectomy (SG), and duodenal switch (DS). Data of the subgroup of patients with BMI 25-30 were detailed separately, as there is a paucity of such data in the literature. Postoperative composite adverse outcome was defined as presence of any of 16 major adverse events. Results: The mean BMI of the 1,003 patients who met the inclusion criteria was 33.5 1.6 kg/m. Fortysix patients had a BMIo30 kg/m. Forty percent were taking insulin and 60% were on oral hypoglycemic medications. The most prevalent baseline comorbidities were hypertension (77.6%) and cardiac diseases (9.1%). Surgical procedures included RYGB (57.2%), AGB (22.6%), SG (18.8%), and DS (1.3%). The mean operative time and length of hospital stay were 110.3 51.6 min and 2.0 1.7 days, respectively. Incidence of all individual complications was r 0.5% in this cohort except postoperative bleeding, which was 1.6%. Thirty-day postoperative mortality and composite adverse event rates were 0.2% and 4.2%, respectively. Reoperation within 30 days after the primary procedure was necessary in 1.6% of patients (Table 1). Conclusion: A 2-hour surgical procedure requiring a two day hospital stay that is associated with modest early morbidity (4%) and low mortality (0.2%) can lead to remission of a chronic, progressive and disabling disease. Based on these findings, bariatric surgery can be considered a relatively safe option for managing T2D in patients with mild obesity. Further large clinical studies on long-term safety and efficacy outcomes of bariatric surgery in patients with T2D and low BMI are warranted.

Early hospital readmission after bariatric surgery.

With the rise in bariatric procedures being performed nationwide and the growing focus on quality and outcome measures, reducing early hospital readmission (EHR) rates has garnered great clinical interest. The aim of this study was to identify the incidence, reasons, and risk factors for EHR after bariatric surgery. Using American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) dataset (2012-2013), patients with a diagnosis of obesity and body mass index ≥35 who underwent bariatric surgery were identified. EHR was defined as at least one hospitalization within 30days of bariatric procedure. The association between readmission and patient factors was assessed using multivariable logistic regression analysis. In addition, reasons for readmission were sought. A total of 36,042 patients were identified. The overall EHR rate was 4.70% [laparoscopic (lap) adjustable band 1.87%, lap gastric bypass (GBP) 5.94%, open GBP 7.86%, and sleeve gastrectomy 3.73%], and it occurred at the median of 11days postoperatively. The rate of EHR significantly decreased from 2012 to 2013 (5.15 vs. 4.32%, p<0.001). The median age and BMI were 44years and 44.7kg/m(2), respectively, 78.99% were female, and 70.78% were white. The most common reason for readmission was nausea/vomiting (12.95%), followed by abdominal pain (11.75%) and dehydration (10.54%). On multivariable analysis, factors most strongly associated with readmission were procedure type (lap band: reference; open GBP: OR 3.78, 95% CI 2.47-5.80; lap GBP 3.18, 2.39-4.22; sleeve gastrectomy: 2.03, 1.52-2.71; all p<0.001), steroid use (1.82, 1.33-2.48, p<0.001), pre-discharge complication (1.64, 1.20-2.24, p<0.001), and black population (1.51, 1.34-1.71, p<0.001). Bariatric centers should consider implementing standardized protocols for outpatient monitoring of patients identified to be at high risk of experiencing early readmission, which in turn would decrease overall costs and improve quality of care.

Predictors of hospital readmission following revision total knee arthroplasty.

Hospital readmission is emerging as an important quality measure, yet modifiable predictors of readmission remain unknown. This study was designed to identify risk factors for readmission following revision total knee arthroplasty. The National Surgical Quality Improvement Program dataset was queried to identify patients undergoing revision total knee arthroplasty from 2011 to 2012. Patient demographics, medical co-morbidities, laboratory values, surgical characteristics and surgical outcomes were examined using bivariate and multivariate logistic regression to identify significant predictors for readmission within 30days of discharge. There were 108 readmissions (6.2%) among 1754 patients. Risk factors for readmission included a history of transient ischaemic attack/cerebrovascular accident (OR 3.47; 13 95% CI 1.30, 9.25), female sex (OR 1.75, 95% CI 1.15, 2.68) and general anaesthesia (OR 14 1.74, 95% CI 1.09, 2.79). Hypertension treated with medication (OR 0.61, 95% CI 0.39, 0.96) was associated with a lower risk of readmission. Post-operative complications that were significant predictors of hospital readmission included periprosthetic joint infection (OR 15.09, 95% CI 5.57, 40.91), superficial wound infection (OR 16.57, 95% CI 5.82, 47.22) and deep venous thrombosis (OR 8.59, 95% CI 2.36, 31.24). The preferred use of neuraxial anaesthesia and coordinated discharge planning in patients with a history of transient ischaemic attack/cerebrovascular accident may reduce the risk of readmission following discharge after revision total knee arthroplasty. Additionally, patients with post-operative infections and deep venous thrombosis following these procedures can benefit from close observation in the first weeks following discharge to minimize the likelihood of readmission. III.