ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS 329

were female. 55.3% of patients were initially diagnosed as inpa-tients, of whom 15.0% underwent surgery during that hospitaliza-tion. The index visit was the only encounter for diverticulitis in80.4% of individuals, while 19.6% of patients had at least two,6.5% had at least three, and 5.7% had four or more primary encoun-ters for diverticulitis. Of patients who did not have surgery at theirincident visit, the majority (87.9%) did not go on to have surgery(Figure). Inpatient treatment for diverticulitis did not predict thesubsequent need for surgery, and of all patients who did have sur-gery, 97.4% underwent surgery during the first year after incidentdiagnosis. Conclusions: The majority of elderly patients newly di-agnosed with diverticulitis neither went on to require surgery, norexperienced recurrent episodes. Most surgical treatment of divertic-ulitis occurred during the hospitalization associated with the indexdiagnosis. The apparent benign course of this disease in older pa-tients suggest that a more conservative approach to the manage-ment of this population may be warranted.

47.3. Frailty in Patients Referred For Liver Transplant Eval-uation: An Independent Domain Associated With De-creased Survival. M. E. Harris, C. Harbaugh, J. F.Friedman, S. A. Lahidji, A. N. Roberts, M. J. Englesbe, C. J.Sonnenday; University of Michigan, Ann Arbor, MI

Introduction: The current allocation system for liver transplanta-tion is based upon the Model for End Stage Liver Disease (MELD)score, which predicts waitlist mortality based upon severity of under-lying liver disease. Current efforts in the transplant community seekto identify patients with adequate physiologic reserve to withstandliver transplantation and enjoy acceptable post-transplant outcomes,factors that are not well predicted by MELD. In gerontology, the con-cept of frailty has been espoused as a measure to predict adversehealth outcomes. However, frailty has not been specifically studiedamong candidates for liver transplantation. In this study, we reportour initial experience in the use of frailty as a method to evaluate pa-tients being considered for liver transplantation. Methods: A pro-spective cohort study of adult patients referred for liver transplantevaluation. Each patient was assessed using a validated tool forfrailty measurement, measuring grip strength, walking speed, unin-tentional weight loss, exhaustion, and physical activity, as originallydescribed by Fried (JGeron 2001). Each subject was assigned a frailtyscore 0 to 5, with higher scores reflecting greater frailty. Measures ofseverity of liver disease and comorbid diagnoses (DM, CAD, HTN,PVOD, COPD, CHF, tobacco use, previous cancer) were collectedand compared to frailty in their ability to predict pre-transplant sur-vival. Results: To date, the study population includes 166 patientswith ESLD referred for liver transplant evaluation, with amedian fol-low-up of 4 months (range 2-13). Frailty scores were normally distrib-uted across this population of chronically ill patients. Frailty was notwell correlated with either MELD score (r¼0.11, P¼0.18) or comorbiddisease (r¼0.02, P¼0.76), suggesting that frailty is a unique domain ofpatient risk.High frailty patients (score 3-5, n¼67) experiencedmark-edly decreased short-term survival (3 month survival 37% vs. 67%,P¼0.01) when compared to low frailty patients (score 0-2, N¼99).By Cox proportional hazards analysis, high frailty proved to a morepowerful independent predictor of adjusted mortality (HR¼2.7,P¼0.02) than any other covariate including MELD score. The combi-nation of high frailty and MELD>15 predicted particularly poor sur-vival (P<0.001), as displayed in the Figure. Conclusions: Amongpatients referred for liver transplantation, frailty appears to vary ina normally distributed manner, and appears independent of severityof liver disease or comorbidity. High frailty independently predictspoor pre-transplant survival, and its effect appears magnified in pa-tients with more severe liver disease. Prospective study is ongoingto correlate frailty with post-transplant outcomes among liver trans-plant recipients.

47.4. Positive Predictive Value Of The AHRQ Patient SafetyIndicator Postoperative Sepsis: Implications For Prac-tice And Policy. M. Cevasco,1,2 A. M. Borzecki,3,4,6 A. K.Rosen,3,4 Q. Chen,3 P. A. Zrelak,7 P. S. Romano,8 M. Shin,3

K. M. Itani1,5,6; 1VA Boston Healthcare System, WestRoxbury, MA; 2Brigham and Women’s Hospital, Boston, MA;3Center for Organization, Leadership, and ManagementResearch, Jamaica Plain, MA; 4Boston University School ofPublic Health, Boston, MA; 5Harvard Medical School, Boston,MA; 6Boston University School of Medicine, Boston, MA;7Center for Healthcare Policy and Research, Sacramento, CA;8Division of General Medicine and Center for HealthcarePolicy and Research, Davis, CA

Introduction: Patient Safety Indicator (PSI) 13, or ‘‘PostoperativeSepsis,’’ of the Agency for Healthcare Quality and Research(AHRQ), was recently adopted as a consensus standard for qualityof care by the Centers for Medicare and Medicaid (CMS). We soughtto examine its positive predictive value (PPV) by determining howwell it identifies true cases of postoperative sepsis.Methods:Two ret-rospective cross-sectional studies of hospitalization records that metPSI 13 criteria were conducted, one within the Veterans Administra-tion (VA) Hospitals from fiscal year (FY) 2003 to FY 2007; and onewithin private sector hospitals between October 1, 2005 and March31, 2007. Trained abstractors reviewedmedical records from each da-tabase using standardized abstraction instruments. We determinedthe PPV of the indicator and performed descriptive analysis of cases.Results:Of the 112 cases flagged and reviewedwithin the VA system,59 were true events of postoperative sepsis, yielding a PPV of 53%(95% CI 42-64%). Within the private sector system, of 164 flaggedand reviewed cases, 67 were true cases of postoperative sepsis, yield-ing a PPV of 41% (95% CI 28-54%). False positives were due to infec-tions that were present on admission; urgent or emergent cases; noclinical diagnosis of sepsis; or other coding limitations such as non-specific shock in postoperative patients. Conclusions: PSI 13 hasa poor predictive ability to identify true cases of postoperative sepsisin both the VA and private sectors. Inherent coding limitations werethe primary reason for false positives. As it currently stands, theuse of PSI 13 for hospital reporting is premature.

47.5. Is Subtotal Thyroidectomy a Cost-Effective Treatmentfor Graves’ Disease? K. A. Zanocco, M. Heller, D. Elaraj, C.Sturgeon; Northwestern University Department of Surgery,Chicago, IL

Introduction: The ideal treatment for Graves’ disease (GD) shouldresult in rapid and durable resolution of hyperthyroidism with

ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS330

a low complication rate and cost. The three commonly-available treat-ment options for GD are antithyroid drugs (ATDs), radioactive iodine(RAI), and thyroidectomy. Thyroidectomy is the least utilized optionin the United States, and when performed is usually a near-total thy-roidectomy. Subtotal thyroidectomy is rarely used due to concernsover cost, high recurrence rate, and low postoperative euthyroidrate.We hypothesized that, under the proper conditions, subtotal thy-roidectomy would be a cost-effective treatment for patients with GDwhen compared to RAI, ATD, or total thyroidectomy. The specificaims of this study were to characterize the conditions which would fa-vor subtotal thyroidectomy for GD. Methods: Cost-effectivenessanalysis was performed using a Markov transition-state model tocompare operative versus medical treatment strategies for GD. A de-cision model was developed based on a standardized reference case ofa 30 year old patient with GD. Treatment outcomes and their proba-bilities were identified based on literature review. Reference case as-sumptions for the outcomes of subtotal thyroidectomy included a 32%rate of hypothyroidism, 60% euthyroidism, and 8% recurrent GDbased on meta analysis data. Costs were estimated using Medicarecharge and reimbursement data and the Nationwide Inpatient Sam-ple. Outcomes wereweighted using QOLutility factors, yielding qual-ity-adjusted life years (QALYs) as a measure of effectiveness. Allfuture costs and QALYs realized were assigned a 3% discount rate.Univariate and multivariate sensitivity analysis and Monte Carlosimulation were used to examine the uncertainty of costs, probabili-ties, and utility estimates in the model. Results: The subtotal thy-roidectomy strategy produced 25.783 QALYs. The incremental cost-effectiveness ratio was $29,847 per QALY gained, reflecting a gainof 0.091 QALYs at an additional cost of $2,710 compared to RAI.The total thyroidectomy strategy yielded fewer QALYs (25.611)than subtotal thyroidectomy or RAI. The ATD strategy was the leasteffective, producing an expected 25.591 QALYs. Sensitivity analysisdemonstrated that subtotal thyroidectomy was cost-effective com-pared to RAI only if the initial postoperative euthyroid rate wasgreater than 50%, the recurrence ratewas less than 18%, and the totalcost of surgery was less than $6,700. Monte Carlo simulation showedthe subtotal thyroidectomy strategy to be optimal in 792/1,000 hypo-thetical GD patients. Conclusions: This study demonstrates thatsubtotal thyroidectomy can be a cost-effective treatment for GD.How-ever, a thyroidectomy technique with an initial postoperative euthy-roid rate greater than 50% and a recurrence rate of less than 18%was necessary.

QUICKSHOT SESSION: THURSDAY, FEB 3, 20113:30 - 5:30 PM

CARDIOTHORACIC 2: THORACIC&VASCULAR

48.1. Transition FromOpen Pulmonary Lobectomy To Thora-coscopic Lobectomy As The Standard Of Care For EarlyStage Lung Cancer Does Not Detrimentally Affect Resi-dent Operative Experience. D. T. Cooke, A. P. Mahfoozi,V. Kuderer, J. Young, R. F. Calhoun; University of California,Davis Medical Center, Division of Cardiothoracic Surgery,Sacramento, CA

48.2. The Value Of Veterans Affairs Hospitals In Cardiotho-racic Surgical Training. F. G. Bakaeen,1 E. H. Stephens,1

D. Chu,1 J. S. Coselli,1 W. L. Holman,2 A. A. Vaporciyan,3

B. L. Cmolik,4 W. H. Merril,5 F. L. Grover6; 1Baylor College ofMedicine and the Michael E. DeBakey VA Medical Center,Houston, TX; 2University of Alabama At Birmingham,Birmingham, AL; 3The University of Texas MD AndersonCancer Center, Houston, TC; 4Louis Stokes Cleveland VAMedical Center, Cleveland, OH; 5University of Mississippi,

Division of Cardiothoracic Surgery, Jackson, MS; 6Universityof Colorado Denver, Aurora, CO

48.3. The Evaluation of a Novel Computed Tomography Vol-ume Index Score (CTVI) for Pulmonary Contusion toAccuratelyPredictOutcomes inPatientsWithBlunt Tho-racic Injury. A. Strumwasser, E. Chu, E. Cureton, R. Kwan, K.Dozier, L. Yeung, E. Miraflor, J. Sadjadi, G. Victorino;UCSF-East Bay Department of Surgery, Oakland, CA

48.4. Three-dimensional Computed Tomographic Analysis ofBronchial Arteries for Preoperative Simulation ofEsophageal Cancer Surgery. T. Wada,1 T. Oyama,1 R.Nakamura,1 N. Wada,1 M. Jinzaki,2 Y. Saikawa,1 S.Kuribayashi,2 H. Takeuchi,1 T. Takahashi,1 Y. Kitagawa1;1Department of Surgery, School of Medicine, Keio University,Shinjuku-ku, Tokyo; 2Department of Radiology, KeioUniversity School of Medicine, Shinjuku-ku, Tokyo

48.5. Predictors Of In-hospital Complications After Pericar-diectomy: A Nationwide Outcomes Study. R. R.Gopaldas,1 A. K. Tharakan,1 T. K. Dao,2 J. G. Markley,1 N. R.Caron1; 1University of Missouri-Columbia School of Medicine,Columbia, MO; 2University of Houston, Houston, TX

48.6. Timely Tracheobronchial Stenting Provides SignificantCost-Savings In Malignant Central Airway Obstruction.K. Park, S. S. Razi, G. Schwartz, S. Belsley, G. Todd, C. P.Connery, F. Y. Bhora; St. Luke’s - Roosevelt Hospital Center,Columbia University College of Physicians & Surgeons, NewYork, NY

48.7. Predictors Of Limited Resection For Early Stage Non-Small Cell Lung Cancer. S. E. Billmeier,1 J. Z. Ayanian,1,3

A. M. Zaslavsky,3 D. R. Nerenz,2 M. T. Jaklitsch,1 S. O.Rogers1; 1Brigham and Women’s Hospital, Boston, MA;2Center for Health Services Research, Henry Ford HealthSystem, Detroid, MI; 3Harvard Medical School Department ofHealth Care Policy, Boston, MA

48.8. Pemetrexed (Alimta) Cytotoxicity In An In Vitro Chemo-response Assay For Non-Small Cell Lung Carcinoma(NSCLC). M. J. Schuchert,1 R. J. Landreneau,1 R. J.Cerfolio,2 J. D. Luketich,1 R. J. McKenna,3 C. B. Fuller,3 S. L.Suchy,4 S. L. Brower,4 P. R. Ervin4; 1Department ofCardiothoracic Surgery; University of Pittsburgh MedicalCenter, Pittsburgh, PA; 2Cardiothoracic Surgery; Universityof Alabama At Birmingham, Birmingham, AL; 3Cedars-SinaiMedical Center, Los Angeles, CA; 4Precision Therapeutics,Inc., Pittsburgh, PA

48.9. Cellular Proliferation And Lung Recruitment Of CD133+/FlK-1+ Cells In The Initial Phase Of Post-Pneumonec-tomy Compensatory Lung Growth (PPCLG). N.Ghobril,1,2 F. Lim,1,2 S. Lang,1,2 H. N. Jones,1,2 L. D. Le,1,2

S. G. Keswan,1,2 M. Habli,1,2 T. M. Crombleholme1,2; 1TheCenter for Molecular Fetal Therapy, Cincinnati Children’sHospital Medical Center, Cincinnati, Ohio; 2Division ofPediatric General & Thoracic Surgery, Cincinnati Children’sHospital Medical Center, Cincinnati, Ohio