gA[s

CmTid

OtPJU

IfTti

Mm8awsi

Rw2aBht(29c

CbPf

SRisSBA

Ig

ptsi

Mwhm

Ra((3so8(brp

Cimtsrso

EstAEESD

Ineaotv

Mapmav

S104 Surgical Forum Abstracts J Am Coll Surg

ency referral, adjusting for procedure type, gender, and age, found thatfrican American race (odds ratio [OR] � 2.1; 95% confidence interval

CI], 1.9, 2.5) and uninsured status (OR � 3.2; 95% CI, 2.5, 4.1) wereignificant independent variables.

ONCLUSIONS: Compared with Caucasians, African Americans areore likely to access surgical care through the emergency department.hese findings are independent of insurance status and are important to

nform the discourse on health care reform. Insurance coverage aloneoes not ensure appropriate access for all patients requiring surgery.

besity is a risk factor for nosocomial infections inhe trauma populationablo E Serrano MD, MPH, MSBS, Sadik Khuder PhD, MPH,ohn Fath MD, MPHniversity of Toledo, Toledo, OH

NTRODUCTION: Obesity, like multiple trauma, results in an in-lammatory condition that leads to an immunodeficiency state.hus, obese trauma patients are at an extremely high risk of infec-

ion. Despite this risk, studies to date have not defined obesity as anndependent risk factor for infection in trauma patients.

ETHODS: Retrospective data were collected on 645 patients ad-itted to a university-based Level 1 trauma center during an

-month period. Obesity was defined as a BMI � 30. Outcomesnalyzed included total infection rate, UTI, pneumonia, septicemia,ound and Clostridium difficile infections. Multiple logistic regres-

ion was used to evaluate the contribution of each BMI category tonfection while adjusting for comorbidities, age, sex, and ISS.

ESULTS: Obesity prevalence was 31.5%. The majority of patientsere men (62% vs 38%). Women were more likely to be obese (39% vs7%; p�0.01). The overall rate of nosocomial infections was 9%. Vari-bles independently associated with increased risk of infections wereMI, age, ISS, and multiple comorbidities. The larger the BMI, theigher the percentage of occurrence and the risk of nosocomial infec-ions: BMI � 25: 6%, BMI 25–29: 8.1%, odds ratio (OR), 1.103confidence interval [CI]: 0.495–2.456); BMI 30–39: 12.7%, OR:.091 (CI: 0.959–4.562); and BMI � 40: 20%, OR: 3.543 (CI: 1.257–.989). Wound infections (p�0.03) and UTI (p�0.02) were signifi-antly more frequent in obese patients.

ONCLUSIONS: In this retrospective study, obesity was shown toe an independent risk factor for nosocomial infection after trauma.rospective studies would clarify the impact of obesity on recoveryrom traumatic injuries.

tandardization of the fully stapled laparoscopicoux-en-Y gastric bypass for obesity reduces early

mmediate postoperative morbidity and mortality: Aingle center study on 2,606 patientsebastiaan Van Cauwenberge MD, Sebastiaan Lambert MD,runo Dillemans MDZ Sint-Jan AV Hospital, Brugge, Flanders, Belgium

NTRODUCTION: Various techniques of laparoscopic Roux-en-Y

astric bypass have been described. We completely standardized this t

rocedure to minimize its sometimes substantial morbidity and mor-ality. This study describes our experience with the standardized fullytapled laparoscopic Roux-en-Y gastric bypass (FS-LRYGB) and itsnfluence on the 30-day morbidity and mortality.

ETHODS: We retrospectively analyzed 2,606 patients who under-ent FS-LRYGB from May 2004 to August 2008. Operative time,ospital stay and readmission, reoperation, and 30-day morbidity/ortality rates were then calculated.

ESULTS: There were 539 male and 2,067 female patients. Meange was 39.2 years (range 14–73 years), mean BMI 41.44 kg/m2

range, 23–75.5 kg/m2). The mean hospital stay was 3.35 daysrange, 2–71 days). Mean total operative time was 63 minutes (range,5–150 minutes). One patient died of pneumonia within 30 days ofurgery (0.04%). One hundred fifty one (5.8%) patients had post-perative complications as follows: gastrointestinal hemorrhage (n �9, 3.42%), intestinal obstruction (n � 9, 0.35%), anastomotic leakn � 5, 0.19%), and others (n � 47, 1,80%). In 66 patients, theleeding resolved without any surgical reintervention. One hemor-hage resulted in hypovolemic shock with subsequent renal and he-atic failure.

ONCLUSIONS: The systematic approach and the full standard-zation of the FS-LRYGB procedure contribute highly to the very low

ortality and the low morbidity rates in our institution. Gastroin-estinal bleeding appears to be the commonest complication, but iself-limiting in the majority of cases. Our approach also significantlyeduces operative time and turns the technically demanding laparo-copic Roux-en-Y gastric bypass procedure into an easy reproducibleperation, effective for training.

ndovenous ablation versus saphenous veintripping: A profit, cost, and efficiency analysis inhe United Statesbhishek Chatterjee MD, MBA, Lilian Chen BS,douard Aboian MD, Shirley A Montagne MBA,va M Rzucidlo MD, FACS,amuel RG Finlayson MD, MPH, FACSartmouth Hitchcock Medical Center, Lebanon, NH

NTRODUCTION: Given increasing health care spending, there is aeed to evaluate new surgical technology with regards to cost andfficiency. Scarce cost analysis literature exists between endovenousblation (EA) and saphenous vein stripping (SVS) for the treatmentf varicose veins in the United States. Our goal is to perform a profit,otal cost, and efficiency analysis for each method in the treatment ofaricose veins.

ETHODS: A financial database search was done for all 2008 EAnd SVS cases performed on one extremity at our tertiary care hos-ital. We compared each method for direct and indirect costs, profitargin, and operating time. We compared efficiency by calculating

n opportunity cost multiplier which represented additional varicoseein procedures that could have otherwise been performed in the

ime spent by the less efficient method.

R�gbcEt(

T

Y

C

N

D

I

T

T

P

P

O

O

O

Capi

CcNBWU

Icact(

MsWTi4epKvp

Rr

rytfrsl

CsrlWt

R1DESU

IcCrmc

Msfirm

Rwptadwjo7m

Ccasct

S105Vol. 209, No. 3S, September 2009 Surgical Forum Abstracts

ESULTS: Total costs (Table 1) were similar for each method in the10 and 10–20 phlebectomy stab incision groups. The profit mar-

in (Table 1) was higher for the EA group independent of the num-er of phlebectomy stab incisions. We calculated the opportunityost multiplier ($21/min) by dividing the profit margin/case of anA with 10–20 stab incisions by its operating time. The SVS method

ended to be less efficient and, hence, incurred an opportunity costTable 1) that made it more expensive to perform than EA.

able 1

ear 2008

< 10 PhlebectomyStab Incisions

10–20 PhlebectomyStab Incisions

20 PhlebectomyStab Incisions

EndovenousAblation

SaphenousVein

StrippingEndovenous

Ablation

SaphenousVein

StrippingEndovenous

Ablation

SaphenousVein

Stripping

ases (#) 8 6 18 22 9 4

et revenue ($) 75,530 40,191 178,446 18,444 110,854 38,440

irect costs ($) 38,244 26,924 96,226 110,145 54,544 19,437

ndirect costs ($) 11,790 10,451 30,817 43,554 16,883 7,520

otal costs ($) 50,034 37,375 127,043 153,699 71,427 26,957

otal costs/cases ($) 6,254 6,229 7,058 6,986 7,936 6,739

rofit margin ($) 25,496 2,816 51,403 34,745 39,427 11,483

rofit margin/cases ($) 3,187 469 2,856 1,579 4,381 2,871

R time (min) 116 149 136 164 143 155

R time difference 33 28 12

pportunitycost/cases ($)

693 588 252

ONCLUSIONS: Endovenous ablation may be more profitablend efficient when compared with traditional saphenous vein strip-ing for the treatment of varicose veins. Future cost-effective analysiss needed to determine societal benefits.

onditional survival after wide-local excision ofutaneous melanoma: Population-based analysisatasha M Rueth MD, Shawn S Groth MD,eth Virnig PhD, MPH, Elizabeth B Habermann PhD, MPH,addah Al-Refaie MD, Todd Tuttle MD, MPH, FACSniversity of Minnesota, Minneapolis, MN

NTRODUCTION: Because melanoma survival curves are nonlinear,urrent estimates may not accurately predict long-term outcomesfter a certain survival period. Our objective is to examine whetheronditional survival data are a more useful tool in predicting long-erm melanoma survival in patients following wide local excisionWLE).

ETHODS: We used the Surveillance, Epidemiology, and End Re-ults database (1992–2005) to identify patients who underwent

LE for nonmetastatic melanoma. We included patients with T2,3, or T4 disease and with known nodal status. Patients were strat-

fied into low-risk (T2–3N0M0) and high-risk (T4N0M0 or T2–N1M0) categories. We defined conditional survival as time-specificstimates conditioned on living to a certain point in follow-up anderformed unadjusted conditional survival analysis using theaplan-Meier method. We adjusted for potential confounding co-ariates using a Cox proportional hazards regression model (al-ha�0.05).

ESULTS: 8,759 patients met inclusion criteria. At diagnosis, low-

isk patients had a significantly improved 10-year survival rate (low- p

isk, 82%; high-risk, 49%; p�0.001). On conditional survival anal-sis, however, this difference holds only until 8 years after WLE; athat point, conditional survival rates are no longer significantly dif-erent (p�0.51) between low- and high-risk groups (adjusted low-isk 97.6%, adjusted high-risk 97.0%; p�0.73). Further cancer-pecific survival beyond 8 years is not predicted by age, gender,ocation, or ulceration.

ONCLUSIONS: As survival time increases, melanoma conditionalurvival data are discordant with traditionally used, AJCC-basedates. These findings can have important implications for surveil-ance and counseling, as high-risk melanoma survivors 8 years after

LE may require no more intensive surveillance than low-risk pa-ients.

eadmission after colectomy for cancer predicts-year mortalityavid Yu Greenblatt MD, MSPH, Maureen A Smith MPH, PhD,rin S O’Connor MD, Noelle K LoConte, Jinn-Ing Liou MS, MBA,haron M Weber MD, FACSniversity of Wisconsin, Madison, WI

NTRODUCTION: Early hospital readmission is a common andostly problem in the Medicare population. Beginning in 2009, theenters for Medicaid and Medicare Services will mandate hospital

eporting of disease-specific readmission rates. We sought to deter-ine the rate and predictors of readmission after colectomy for can-

er, as well as the association between readmission and mortality.

ETHODS: Medicare beneficiaries who underwent colectomy fortage I–III colon adenocarcinoma from 1992–2002 were identifiedrom the SEER-Medicare database. Multivariate logistic regressiondentified predictors of early readmission and 1-year mortality. Oddsatios were adjusted for multiple factors, including measures of co-orbidity, socioeconomic status, and disease severity.

ESULTS: Of 42,375 patients who were discharged, 4,662 (11%)ere readmitted within 30 days. The most common causes of rehos-italization were ileus/obstruction and infection. Significant predic-ors of readmission included male gender, comorbidity, emergencydmission, prolonged hospital stay, blood transfusion, ostomy, andischarge to nursing home. Readmission was inversely associatedith hospital procedure volume, but not surgeon volume. After ad-

usting for potential confounding variables, the predicted probabilityf 1-year mortality was 16% for readmitted patients, compared with% for those not readmitted (p-value �0.001). The difference inortality was significant for all stages of cancer.

ONCLUSIONS: Early readmission after colectomy for cancer isommon and due in part to modifiable factors. There is a remarkablessociation between readmission and 1-year mortality. Early readmis-ion is therefore an important quality-of-care indicator for colonancer surgery. These findings may facilitate the development ofargeted interventions that will decrease readmissions and improve

atient outcomes.