QS184. Rural Vs. Urban Trauma: Demographic Influences on Autopsy Rates

340 ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS contractile state. Though it is not feasible to hold parameters constant in critically ill patients, we believe that the echocardiographic assessment of left ventricular function in groups of patients, stratified for disease pathogenesis, does reflect reversibity/irreversibility of myocardial contractile depression. Conclusion: Left ventricular dysfunction is reversible in 1 out of 3 patients with cocaine toxicity, sepsis, and fluid overload. Our observations provide encouraging data to suggest that documented myocardial depression in critically ill patients deserves aggressive intervention with the likely potential for hemodynamic improvement. References: 1. Rudiger A, Singer M, “Mechanisms of sepsis-induced cardiac dysfunction”. Crit. Care Med. 2007 Jun; 35(6):1599 – 608. QS182. THE IMPACT OF THE 80 HOUR WORK WEEK ON MORTALITY IN TRAUMA PATIENTS: AN ANALYSIS OF THE NATIONAL TRAUMA DATA BANK. Catherine A. Morrison, Matthew M. Wyatt, Matthew M. Carrick; Baylor College of Medicine, Houston, TX Introduction: The implementation of the 80 hour work week (80HrWW) restrictions implemented by the Accreditation Council for Graduate Medical Education (ACGME) in July 2003 were intended, in part, to improve patient outcomes by reducing fatigue-related resident errors. Although concerns were raised regarding the possibility for increased error due to decreased continuity of patient care, recent studies have shown no significant change in mortality or complication rates since the onset of these new restrictions. This study is the first to examine the effects of the 80 hour work week on mortality in trauma patients on a national level. Methods: Data were obtained from the National Trauma Data Bank (version 6.2) from 1994-2005. Data were then divided into two groups: “Pre80HrWW” (2001-2002) and “Post-80HrWW (2004 - 2005). Because the ACGME’s guidelines were implemented mid-year in 2003, and because the National Trauma Data Bank classifies admission date only by year, all patients admitted during 2003 were excluded from the analysis. Information regarding demographics, hospital type (university versus non-teaching), and geographical region (Northeast, South, Midwest, West) was collected. Our primary outcome measure was mortality. Age, hospital type and geographical location were analyzed as independent variables using logistic regression. Results: The overall mortality rate decreased from 4.64% in the Pre-80HrWW to 4.46% in the Post-80HrWW (p⬍0.0001). A similar trend towards decreased mortality was also seen in the geographical regions we examined. Of particular interest were the differences in outcomes observed in academic versus non-academic institutions. In university hospitals, the mortality decreased from 5.16% to 5.03% (p⫽0.03); whereas in non-teaching hospitals, mortality actually increased from 3.37% to 3.85% (p⬍0.001). Discussion: A number of papers have shown decreased performance in technical skills and clinical judgment in sleep-deprived physicians. Many have suggested that decreased resident work hours would likely result in improved patient outcomes. However, critics of the ACGME’s policy claim that decreased continuity of care and reduced exposure to patients has the potential to cause more resident errors and thus negatively affect patient outcomes. Few papers exist which specifically examine patient outcomes within the field of surgery, and even fewer papers investigate the effects of the 80 hour work week on trauma patients in particular. This large retrospective analysis confirms that the 80 hour work week has not negatively affected patient outcomes and has in fact significantly decreased mortality for trauma patients in university hospitals nationwide. Although the absolute decrease in mortality rate is small, when accounting for a trauma population of millions of patients, this translates into thousands of lives saved per year. QS183. GERIATRIC PATIENTS WITH HEAD INJURIES AND SPLENIC INJURIES: A REVIEW OF THE NATIONAL TRAUMA DATA BANK. William L. Newcomb, William W.

Hope, Thomas M. Schmelzer, Amanda L. Walters, Cynthia L. Lyons, A. Britton Christmas, B. Todd Heniford, Ronald F. Sing; Carolinas Medical Center, Charlotte, NC Background: Non-operative management of splenic injuries in hemodynamically stable geriatric patients is preferred if the patient is stable and can be closely monitored. Head injured patients, however, have an unreliable abdominal exam, and even brief periods of instability in geriatric patients are not tolerated well and lead to increased morbidity. Methods: The National Trauma Data Bank® (NTDB) was reviewed from 2001-2005 for patients ageⱖ65 years with both splenic injuries and head injuries. Patients were excluded for systolic blood pressure ⬍ 90 on arrival and an Abbreviated Injury Severity (AIS) abdominal score ⬍ 2. Patients who did not undergo splenectomy or splenorrhaphy ⱖ 2 hours after diagnosis were considered non-operative successes. Results: In the four-year period, 351 patients were identified; 97% were managed non-operatively. The non-operative success rate was 91%. Systolic blood pressure on arrival (116 vs. 137, p⫽0.0110), base deficit on arrival (⫺11 vs. ⫺4.1, p⫽0.0127), and Revised Trauma Score (4.3 vs. 6.6, p⫽0.0635) were lower for patients managed operatively. No difference in Injury Severity Score (ISS) was found for operative vs. non-operative patients or for non-operative successes vs. failures. Mortality was higher in the operative group (82% vs. 31%, p⫽0.0009). No differences in mortality, hospital length of stay (LOS), intensive care unit (ICU) days, or hospital charges were seen for patients who failed or succeeded non-operative management. Conclusion: Head injured geriatric patients with splenic injuries can be managed non-operatively in 97% of cases with no difference in mortality, LOS, ICU days, or hospital charges between those who succeed or fail non-operative management. Increased mortality is observed in patients managed operatively, but is associated with worse hemodynamic parameters on arrival. Poorly tolerated hemodynamic instability in the geriatric population underscores the importance of close monitoring for nonoperative management. QS184. RURAL VS. URBAN TRAUMA: DEMOGRAPHIC INFLUENCES ON AUTOPSY RATES. Joseph Zabell, Marcus Nashelsky, Timothy Light, Gerald P. Kealey, Timothy Thomsen; University of Iowa Carver College of Medicine, Iowa City, IA Introduction: In order to maintain trauma certification, trauma systems must have a multidisciplinary peer review system to evaluate selective deaths in order to improve trauma care. The autopsy has long been considered the gold standard for quality assurance review. Studies characterizing autopsies have been completed in large urban centers, but there is a paucity of research regarding autopsies at rural trauma centers. This is problematic considering that a majority of trauma deaths occur in rural areas and death rates for unintentional injuries in rural populations are higher than urban populations. Thus, rural trauma centers have differing characteristics warranting further research into the demographic differences between rural and urban trauma patients and the effects on autopsy rates. This is a demographic study of a rural trauma center, University of Iowa Hospitals and Clinics (UIHC), with the goal of identifying differences between rural and urban trauma systems, and the characteristics of trauma patients on which autopsy was performed. Method: 496 trauma deaths were identified from the UIHC trauma registry between January 2002 and May 2007 (231 of which were autopsied) and demographic data (including age, race, length of hospital stay, etc.) regarding these patients was gathered into a database. These patients were compared via Chi-square analysis to those from similar studies completed in urban trauma centers (Miami and Denver). In addition, patients were placed into one of two groups: those who received autopsies, and those who did not. A linear regression model was used to analyze differences between autopsied and non-autopsied trauma patients in the UIHC data set. Results: There were several differences between UIHC and the urban trauma

ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS 341 centers of Miami and Denver: UIHC had lower autopsy rates (46%) than the urban centers (97% and 68% respectively; p⬍0.0001), significantly higher rates of blunt trauma (83% vs. 68% (p⫽0.0002) and 48% (p⬍0.0001)), lower rates of penetrating trauma (7% vs. 11% (p⫽0.0002) and 49% (p⬍0.0001)), a higher mean age of autopsied patient (58 years vs. 50 and 36 years; p⬍0.0001), and a lower percentage of male trauma patients than Denver (65% vs. 79%; p⫽0.0002). Among the UIHC data, autopsied patients differed from non-autopsied patients in that they tended to be younger (mean age 45 years vs. 71 years; p⬍0.0001) and have a shorter median length of hospital stay (1 day vs. 4 days; p⬍0.0001). Additionally, UIHC autopsy rates for patients with blunt trauma were lower than rates for patients with penetrating or burn trauma (42% vs. 67% and 56%; p⫽0.004). Conclusion: UIHC, a rural trauma center, has a number of demographic characteristics that make it unique from urban trauma centers: an older population, lower percentage of male trauma patients, higher rates of blunt trauma, and lower rates of penetrating trauma. All of these factors influenced the lower rate of autopsies completed at UIHC. Among the UIHC population, those less likely to receive autopsy include older patients, those who expired after 48 hours in the hospital, and patients who suffered blunt injuries. The demographics of trauma patients most likely to receive an autopsy tend to correspond with those of an urban trauma population, thus providing a demographic explanation for the variation in autopsy rates among trauma systems. QS185. PREDICTORS OF MORTALITY IN PATIENTS WITH BLUNT DIAPHRAGMATIC RUPTURE. Marc Zerey1, Thomas M. Schmelzer2, Charles S. Joels2, Ronald F. Sing2, Amy E. Lincourt2, Kent W. Kercher2, B. Todd Heniford2; 1 Sansum Clinic, Santa Barbara, CA; 2Carolinas Medical Center, Charlotte, NC Blunt diaphragmatic rupture (BDR) associated with multisystem injuries is often fatal. Our objective was to review our experience, analyze clinical features, and identify significant predictors of mortality. A review of patients sustaining BDR from 1992 to 2005 at a level-I trauma center was performed. Data derived from an institutional prospective trauma registry included demographics, injury severity score (ISS), injury mechanism, associated injuries, preoperative diagnosis, time to operation, complications, and mortality. Chisquare, Student’s t-test, and Spearman’s correlation were used to determine significance (P ⬍ 0.05). BDR was diagnosed in 74 patients over 14 years. There were 42 males and 32 females with a mean age of 40.7 years (range: 15-92 years; SD: 19.2) and average ISS of 27 (range: 4-66; SD: 13). There were a total of 67 left-sided and 7 right-sided BDRs with no significant differences in mechanism of injury, ISS, time to operation, length of stay or mortality. A radiological diagnosis of BDR was made preoperatively in only 55.7% of patients. Most patients (74.3%) had at least one associated intraabdominal injury. Missed BDR occurred in 13/74 patients (17.6%). The absence of an associated intra-abdominal injury strongly correlated with missed BDR (R⫽-0.43, P⫽0.0002). The mortality rate was 21.6% (16/74). No deaths were directly attributable to BDR. The most significant predictors of mortality were: ⬎1 associated intraabdominal injury (P⫽0.02), ISS⬎25 (P⬍0.0001), and initial systolic BPP⬍0.0001). Blunt diaphragmatic rupture is associated with elevated ISS and mortality rate. There is no difference in ISS, mortality, or time to injury diagnosis between left and right-sided blunt diaphragmatic rupture. The diagnosis of diaphragmatic rupture with radiographic imaging is not reliable and missed BDR is common. Death is not directly attributable to the BDR; initial systolic BP, ISS and the presence of multiple abdominal injuries are the important determinants of mortality. QS186. DOES HEMODYNAMIC VARIABILITY PREDICT OUTCOME IN SHOCK? Thomas Street1, Matthew Barton1, Robert Vogel2, Jonathan Whitten1, Christopher Stout1,

Dennis Ashley1, Julie Wynne1, William Thompson1, Joseph Van De Water1; 1Mercer University School of Medicine, Macon, GA; 2Georgia Southern University, Statesboro, GA Introduction: Decreased variability of heart rate on a beat to beat basis has been shown to be a reliable predictor of adverse patient outcome in shock resuscitation. Unfortunately, technology is not available in the standard Intensive Care Unit (ICU) to assess hemodynamic parameters at this rate and in such large volumes. Hypothesis: Measurements taken at 5 minute intervals using available and non-invasive equipment, including impedance cardiography (ICG), can be used to predict patient outcome. Method: Patients selected for the study were classified as shock patients if they were 1) septic and/or 2) hemodynamically unstable. Readings were taken every 5 min for 1 hour. Parameters recorded included: heart rate (HR), systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and cardiac index (CI). Endpoints for the study were 1) Non Survivor: death or discharge in a vegetative state or 2) Survivor: discharge from ICU to the ward. There were a total of 28 patients studied. The data were analyzed using mixed-effects regression models (MRM). Results: The parameters MAP, DBP, and HR had greater variability with respect to trend over time for the non-survivor patients.

Parameter CI DBP HR MAP SBP

Variance Non-Survivor

Variance Survivor

p-value

0.000005 0.0076 0.005 0.008 0.0154

0.00005 0.0011 0.0009 0.001 0.0055

1 0.0025 0.0068 0.0012 0.0946

Conclusion: Using non-invasive technology and standard ICU protocols, adverse outcomes can be predicted on the basis of increased variability in MAP, DBP, and HR. These results taken in a different fashion using a standard ICU protocol seem to contradict recently published reports attributing increased variability and bad outcomes to deterioration of the autonomic nervous system or “autonomic uncoupling”. QS187. MEETING THE CHALLENGE OF CONDUCTING A CLINICAL TRIAL IN TRAUMA UTILIZING A RESEARCH ASSISTANT MODEL. Gail Wainwright, Nathaniel McQuay, Jr.; St Luke’s Hospital & University of Pennsylvania Trauma Network, Bethlehem, PA Background: Conducting clinical trials in the field of Trauma presents many challenges. One of these challenges is the possibility of patient enrollment at any time. Enrollment often occurs after hours during which resources are limited. Objective: A Research Assistant (RA) model at a Level I Regional Resource Trauma Center was designed to support participation and to facilitate enrollment in a national multicenter clinical trial. Methods: Sixteen staff clinicians were chosen by predetermined criteria. An information session and four training sessions were implemented. Upon completion of the training session all research assistants were scheduled for call via an internet based calendar. A response protocol included a ten minutes response by telephone and an arrival time of 30 minutes. Variables assessed via retrospective evaluation of case report forms (CRF) were protocol violations, availability, after hour enrollment and successful completion of enrollments. After hours was defined as enrollment occurring during weekends, holidays and between 5pm to 8am Monday through Friday. Results: Sixteen clinicians (14 RN, 1PCT, 1 EMT) made up the RA team. There were a total of 13 patients enrolled over an 18 month period. Seven of 13 (41%) patient enrollments were managed by a RA. Eight of 13 (61%) patients were enrolled during after hours. On two occasions, the research assis-