53.06 Intrahepatic Cholangiocarcinoma Tumor Burden to Predict Prognosis Following Resection

Posted on January 31, 2019

J. Cloyd2, F. Bagante2,6, G. Spolverato2, M. Weiss1, S. Alexandrescu3, H. P. Marques4, L. Aldrighetti5, S. K. Maithel7, C. Pulitano8, C. Pulitano8, T. W. Bauer9, F. Shen10, G. A. Poultsides11, O. Soubrane12, G. Martel13, B. G. Koerkamp14, A. Guglielmi6, E. Itaru15, T. M. Pawlik2  1Johns Hopkins Hospital,Surgery,Baltimore, MD, USA 2The Ohio State University Wexner Medical Center,Surgery,Columbus, OH, USA 3Fundeni Clinical Institute, Bucharest,Surgery,Bucarest, -, Romania 4Curry Cabral Hospital,Surgery,Lisbon, -, Portugal 5Ospedale San Raffaele,Surgery,Milan, -, Italy 6University of Verona,Surgery,Verona, -, Italy 7Emory University School Of Medicine,Surgery,Atlanta, GA, USA 8Royal Prince Alfred Hospital,Surgery,Sydney, -, Australia 9University Of Virginia,Surgery,Charlottesville, VA, USA 10Eastern Hepatobiliary Surgery Hospital,Surgery,Shanghai, -, China 11Stanford University,Surgery,Palo Alto, CA, USA 12Beaujon Hospital,Surgery,Clichy, -, France 13University Of Ottawa,Surgery,Ottawa, Ontario, Canada 14Erasmus University Medical Centre,Surgery,Rotterdam, -, Netherlands 15Yokohama City University School of Medicine,Surgery,Yokohama, -, Japan

Introduction: We sought to investigate the impact of total tumor-burden (i.e. size and number) on patient prognosis following resection of ICC.

Methods: Patients who underwent curative-intent resection for ICC at one of the 15 participating institutions between 2005-2016 were identified.

Results:Among 1,278 patients who underwent surgery for ICC, 423 (33.1%) patients had no lymph-node metastasis (N0), while 224 (17.5%) had nodal disease (N1); 631 (49.4%) did not have a lymphadenectomy (Nx). Median tumor size was 6 cm (inter-quartile range [IQR], 4-8). While 1,016 (81.8%) patients had a single ICC, 226 (18.2%) patients had multifocal disease. On multivariable analysis, after adjusting for lymph-node status, tumor size (logarithmic transformation: HR 1.35) and number of ICC (logarithmic transformation: HR 1.51) demonstrated a strong non-linear association with overall survival (OS)(Log-model; Figure 1). Log-model (AUC 0.588) out-performed both tumor size (c-index 0.572) and number of tumors (c-index 0.539) in predicting OS. Among N0 patients, 5-year OS of patients with a single ICC ranged from 80-70% among patients with ICC <3 cm to 50% for patients with ICC >6 cm. Conversely, among N1 patients, 5-year OS of patients with a single ICC ranged from 60-50% for patients with ICC <3 cm to 40-30% for patients with ICC 3-6 cm and 20% for patients with ICC >6 cm.

Conclusion:Tumor size and number of ICC demonstrated a strong non-linear association with survival following resection of ICC. A log-model tumor burden score may be a better tool to estimate prognosis of patients undergoing curative-intent resection of ICC.

 

53.05 Usefulness of Fluorescence Imaging for Laparoscopic Liver Resection and Complex Biliary Surgery

Posted on January 31, 2019

Y. Kawaguchi1,2, Y. Nomura1, M. Nagai1, N. Tanaka1  1Asahi General Hospital,Department Of Surgery,,Asahi, CHIBA, Japan 2the University of Tokyo,Hepato-Biliary-Pancreatic Surgery Division,Bunkyo, TOKYO, Japan

Introduction: A fluorescence imaging technique using indocyanine green (ICG) as a fluorophore has been increasingly used for hepatobiliary surgery, and visualizes liver cancer and the bile duct as fluorescence. However, the usefulness of the technique for laparoscopic liver resection and complex biliary surgery remains unclear. We aimed to evaluate the identification of liver cancer in laparoscopic approach and the visualization of the bile duct during complex biliary surgery in open approach using ICG-fluorescence imaging.

Methods: (1) Visualization of liver cancer was evaluated in 6 patients (13 lesions) who underwent laparoscopic liver resection. As a fluorophore of the technique, ICG was injected intravenously at a dose of 0.5 mg/kg as a routine liver function test within 2 weeks before surgery. (2) Visualization of the bile duct was evaluated in 7 patients who underwent complex biliary surgery. ICG was administered by intrabiliary (IB) injection (0.025 mg/mL) or by intravenous (IV) injection (2.5 mg). The values of fluorescence intensity (FI) of the bile duct and the liver were calculated using a luminance analyzing software.

Results:(1) Of the 13 lesions, there were hepatocellular carcinoma (n=3) and colorectal liver metastasis (n=10). Fluorescence imaging visualized 8 (61.5%) lesions, which were invisible on the surface but were located less than 10 mm from the liver surface (Figure 1A). In contrast, the other 5 were located more than 10 mm from the liver surface and were not visualized as fluorescence. (2) Fluorescence imaging technique with the IB injection method was used for 6 patients with severe inflammation (n=3), abnormal biliary anatomy (n=2), and perforation of the bile duct (n=1). In contrast, the IV injection method was used for 1 patient with abnormal biliary anatomy. When using the IB injection method, the liver did not provide fluorescence as it showed fluorescence using the IV injection method. The fluorescence of the bile duct was clearly visualized on the low FI of the surrounding structures using the IB injection method (Figure1B) compared to fluorescence imaging using the IV injection method (Figure 1C). The median (range) FI ratio of the bile duct to the liver was 19.1 (5.0-67.7) using the IB injection method while it was 1.4 using the IV injection method.

Conclusion:ICG-fluorescence imaging is useful to visualize liver cancers which were not visible from the liver surface during laparoscopic liver resection. The IB injection method provided clear contrast between the bile duct and the surrounding structures compared with the IV injection method. The IB injection method is useful for recognizing the biliary anatomy, especially when biliary drainage tubes were inserted as an intervention of severe biliary tract infection.

 

53.04 Psychotropic Medications May Lead to Altered Presentation for Gastroesophageal Reflux Disease

Posted on January 31, 2019

M. McNally1, P. Belle1, A. D. Jalilvand1, K. A. Perry1  1The Ohio State Wexner Medical Center,Division Of General And Gastrointestinal Surgery,Columbus, OH, USA

Introduction:  Recognition and treatment of gastroesophageal reflux disease (GERD) is largely dependent on self-reported symptoms and quality of life assessment. Underlying psychiatric diagnoses and associated therapies may alter this presentation. As such, the primary goal of this study is to determine if baseline psychotropic medications impact the presentation of GERD, while secondary objectives included comparing disease-specific quality of life and symptom scales. 

Methods:  All patients who presented for elective surgical management of GERD from 2011 to 2016 at a single academic institution (n=359) were reviewed. Demographic data, presenting symptoms, objective work-up, and baseline scores on the gastro-esophageal reflux symptom scale (GERSS) and quality of life (GERD-HRQL) were obtained through chart review. Patients were grouped based on baseline (at the time of intake evaluation) usage of either a serotonin-modulating medication (SMM) or benzodiazepine (BZM) medication, as a proxy for major depression and anxiety disorders. Differences in presenting symptoms, GERSS and GERD-HRQL scores were carried out using Mann-Whitney U, Chi2, or Fisher’s Exact, as appropriate; a p value of <0.05 was considered statistically significant. 

Results: One hundred and sixty-three patients (45.3%) were on psychotropic medications; SMM use accounted for 80% (n=131), followed by BZM (24%, n=62), and 19% were utilizing both SMM/BZM (n=31). Compared to those not on any psychotropic medications, patients on SMM experienced significantly higher rates of chest pain (29% vs 19%, p=0.03), had higher GERD-HRQL scores for dysphagia (3 (1-4) vs 2(0-4), p=0.03) and bloating (4 (2-5) vs 3 (2-4), p=0.05) Patients on BZM presented more often with regurgitation (77 % vs 55%, p <0.005) and trended towards more chest pain (29% vs 19%, p=0.09), while also scoring significantly worse on the GERD-HRQL (33 (25-41) vs 29 (19-36), p=0.02) and GERSS (45.5 (31-54) vs 38 (22-50), p=0.02). Similarly, patients on SMM/BZM also demonstrated higher rates of regurgitation (81% vs 52%,p=0.003) and scores significantly worse on GERD-HRQL (38 (28-43) vs 29 (19-36), p=0.004) and GERSS (48 (31-63) vs 38 (22-50), p=0.02). Compared to the control group, patients on SMM, BZM, or both SMM/BZM demonstrated no difference in motility or pH testing.

Conclusion: Patients on SMM or BZM who present with GERD have higher rates of regurgitation and chest pain and worse scores on disease specific quality of life and symptom scales, Further studies are needed to evaluate how these medications and their associated psychiatric diagnoses impact outcomes after antireflux surgery. 

 

53.03 Perineural Invasion as a Prognostic Factor in Pancreatic Ductal Adenocarcinoma

Posted on January 31, 2019

H. Takahashi1, E. Katsuta1, K. Takabe1  1Roswell Park Cancer Institute,Surgical Oncology,Buffalo, NY, USA

Introduction:

Perineural invasion (PNI) is the process of neoplastic invasion of nerves, and is a morphologic feature observed in various malignancies. It serves as one of the pathologically determined poor prognostic factors along with lymphovascular invasion, and also can be a source of distant metastasis for some tumors. PNI is widely prevalent in patients with pancreatic ductal adenocarcinoma (PDAC). Although the exact mechanism of invasion to nerve still remains unclear, several signaling pathways in tumor microenvironment (TME) have been reported to date. Since PNI is relatively subjective pathological evaluation, there have been some conflicting reports of utility of PNI in the management of PDAC. Therefore, in the present study, we aimed to identify biological factors that are associated with PNI in PDAC using publicly available large data set The Cancer Genome Atlas (TCGA).  

Methods:

Genomic and clinical data of patients with PDAC were obtained from TCGA through cBioportal. Pathological information associated with TCGA was obtained through TIES. Using Kaplan-Meier survival curve and Cox proportional hazards model, clinical and oncologic parameters were analyzed. Gene Set Enrichment Analysis (GSEA) was also conducted between the groups based on PNI status.

Results:

There were 154 patients with PDAC in TCGA. The mean age of the cohort was 65.1 years old, with 83 (54%) patients being male. PNI was positive in 109 patients (71%), negative in 13 (8%), and unknown in 32 (21%). There was no significant difference in overall survival (OS) based on PNI status in TCGA cohort. Median OS was 17.8 months in PNI positive, and 19.9 months in negative (p=0.30). Subsequently, in order to identify the risk factors for OS, univariate analysis was performed with multiple clinical parameters. There was significantly longer OS in the patients who underwent adjuvant radiation therapy or targeted molecular therapy (p=0.002 and p<0.001, respectively). With multivariate analysis, absence of targeted molecular therapy (Hazard Ratio (HR): 4.76, p<0.001), absence of adjuvant radiation (HR: 2.58, p=0.02), and positive PNI (HR: 8.28, p=0.02) were found as independent risk factors of poor prognosis in patients with PDAC.

Furthermore, the PNI positive group was identified to enrich angiogenesis gene set by GSEA (NES: 1.67, p=0.002).

Conclusion:

Positive PNI was one of independent risk factors of poor prognosis in patients with PDAC in this study. It might be due to the associated angiogenesis and possible distant metastasis. 

53.02 Hospital Factors Strongly Influence Robotic Use in General Surgery

Posted on January 31, 2019

C. L. Stewart1, S. Dumitra1, C. Nota1, P. H. Ituarte1, L. Melstrom1, Y. Woo1, G. Singh1, Y. Fong1, H. Nathan2, S. G. Warner1  1City Of Hope National Medical Center,Surgical Oncology/Surgery,Duarte, CA, USA 2University Of Michigan,Hepatobiliary Surgery/Surgery,Ann Arbor, MI, USA

Introduction:   Although the use of robotics in general surgery is increasing in the United States, hospital and patient-level factors driving adoption are sparsely studied.  We hypothesized that general surgeons are more likely to use a robotic surgical platform at hospitals where more urologic and gynecologic robotic surgeries are performed, suggesting that hospital related factors are important for platform choice.

Methods: We queried the Nationwide Inpatient Sample from 2010-2013 for patients who underwent surgery on the gallbladder, pancreas, stomach, spleen, colon, or rectum (general surgery), the prostate or kidney (urologic surgery), and ovaries or uterus (gynecologic surgery).  Operations were classified as robotic if any ICD-9-CM robotic procedure code was used.  Hospitals were grouped into quartiles according to percentage of total volume of urologic or gynecologic surgeries that were performed robotically (0-20%, 21-40%, 41-60%, >60%). Multivariable logistic regression modeling was used to determine independent variables associated with robotic surgery. 

Results:  Survey-weighted results represented 461,368 (47.6%) open, 479,783 (49.5%) laparoscopic, and 27,620 (2.6%) robotic general surgical operations.  For general surgery patients, robotics use increased with each subsequent year studied (5.4% by 2013), and was most commonly performed for rectal surgery (7.0%), on patients with private insurance (3.2%) and higher household income (3.3%, all p<0.001).  Robotic operations were also more frequently performed at urban teaching hospitals (3.4%), compared to rural and non-teaching hospitals (p<0.001).  The odds of a general surgery patient receiving a robotic operation increased directly with increased use in urologic and gynecologic surgery at the hospital (Figure 1, *p<0.001).  General surgery patients treated at a top quartile hospital for robotic urologic surgery had 4 times greater odds of receiving a robotic operation compared to an open or laparoscopic operation (confidence interval 3.0-5.4, p<0.001). This finding was independent of study year, surgical site, insurance type, household income, and hospital type, and also persisted when only comparing laparoscopic to robotic procedures.   

Conclusions:  The use of robotics in general surgery is directly and independently related to its use in urologic and gynecologic surgery at a hospital.  Our study suggests that hospital factors strongly influence robotic use in general surgery. 

53.01 Clinical impact of anatomical resection for HCC treatable with partial resection

Posted on January 31, 2019

T. Gocho1, K. Nakashima1, Y. Shirai1, R. Marukuchi1, J. Yasuda1, H. Shiozaki1, K. Furukawa1, S. Onda1, H. Shiba1, Y. Ishida1, K. Yanaga1  1Jikei University School of Medicine,Department Of Surgery,Tokyo, Japan

Introduction:

Anatomical resection (AR) for hepatocellular carcinoma (HCC) has been reported to have better outcome compared to non-anatomical resection (NR). However, the specific benefit of AR for HCC judged treatable with partial resection remains unclear. The aim of this study is to evaluate the clinical impact of anatomical resection for HCC treatable with partial resection.

Patients and

Methods:
Two-hundred and sixty three patients were treated with primary hepatectomy for HCC between May 1997 and December 2016 at Jikei University Hospital. Of those, 80 patients with solitary tumor which were judged treatable by partial hepatectomy based on preoperative imaging were retrospectively reviewed. We divided such patients into two groups according to the types of resection (anatomical resection (AR) group (n = 28) treated by subsegmentectomy and non-anatomical resection (NR) group (n = 52) treated by partial hepatectomy) and patient factors (age, sex, viral status, ICG R15, Child-Pugh (C-P) grade), tumor factors (size, preoperative AFP and PIVKA-II values and portal vein (PV) invasion), operative factors (operative time, blood loss and blood transfusion) and outcomes (overall (OS) and disease-free (DFS) survival and complications) were assessed.

Results:

The following variables were comparable: sex, HBV infection, HCV infection, ICG 15, C-P grade, Preop. AFP, tumor size, pathological PV invasion, blood loss, blood transfusion and complications. However, AR group had older patients (≥ 65 years) (60% vs. 28.8%, p <0.05), higher preop. PIVKA-II (≥ 100 mAU/ml) (71% vs. 46.2%, p < 0.05) and longer operative time (≥ 360 min) (61% vs. 45%, p < 0.05), which were not independent risk factors related to DFS and OS after primary hepatectomy. There were no statistical difference in 5-year DFS and 5-year OS between AR group and NR group (53% vs. 36%, p = 0.096 and 70% vs. 67%, p = 0.714, respectively). However, for HCC 2 cm or larger, statistically higher 5-year DFS was achieved in AR group as compared with NR group (53% vs. 31%, p = 0.041), while no significant difference was observed in 5-year OS (73% vs. 64%, p = 0.488). Twenty seven of 52 patients in NR group developed intrahepatic recurrence, whose recurrence was in the same subsegment in 9 (33%) and in the other segment in 18 (67%) patients. There was no statistical difference in 5-year OS after recurrence between those with recurrence in the same subsegment and the other subsegment (p = 0.764).

Conclusion:

Anatomical resection (subsegmentectomy) seems to improve DFS of patients with HCC 2 cm or larger in diameter.

52.20 Falls From Tree Stands While Hunting: A Rural Trauma Center’s Experience

Posted on January 31, 2019

L. Keeney1, S. Armen1, S. Allen1  1Penn State University College Of Medicine,Trauma Acute Care And Critical Care Surgery,Hershey, PA, USA

Introduction:
Hunting is a popular sport in the United States, and a fall from a tree stand is the most common mechanism of injury among hunters.  In the past 19 years the rate of tree stand falls among hunters in Pennsylvania has increased. This study aimed to describe typical demographics, injuries sustained and hospital course among hunters in Pennsylvania who suffered a fall from a tree stand.

Methods:
A retrospective descriptive analysis of patients who fell from a tree stand over a 10 year period was conducted using the institutional database at a rural Level 1 trauma center.  Data collected included patient demographics, injury severity score (ISS), vital signs on arrival, injuries sustained, hospital length of stay (LOS), and percent requiring admission to the ICU.

Results:
Of the 57 patients 100% were male with a mean age of 48.7+15.7. Mean ISS was 16.2+8.1 with 95% of patients requiring admission and 25% requiring admission to the ICU.  Vital signs on arrival were relatively normal with a mean systolic blood pressure of 137.1+27.2 mm Hg, pulse of 86.7+13.8, respiration rate of 20.1+4.6, and a median Glasgow Coma Score of 15.  The most frequent injuries include spinal fractures (23%), closed head injury (23%) and extremity fractures (12.3%). 95% of patients required hospital admission and mean LOS was 6.6+5.7 (range: 1-25 days) with nearly 25% requiring admission to the intensive care unit.

Conclusion:
Falls from a tree stand while hunting lead to severe injury that result in prolonged hospitalization and frequently require admission to the ICU.  It is vital to understand common injury patterns and trends in hospital course to raise awareness about tree stand safety.  Hunters and producers of tree stands should understand the magnitude of injury suffered from these falls and aim to improve tree stand safety procedures in order to mitigate hunting related injuries due to falls from tree stands. 
 

52.18 Targeting Provider Beliefs and Practices to Improve Opioid Stewardship

Posted on January 31, 2019

H. F. Thiesset1,2, R. Y. Kim1, V. L. Valentin2, K. Schliep2, L. Gren2, C. A. Porucznik2, L. C. Huang1  1University Of Utah,Department Of Surgery,Salt Lake City, UT, USA 2University Of Utah,Department Of Family And Preventive Medicine,Salt Lake City, UT, USA

Introduction:

In light of the growing opioid epidemic, the over-prescription of opioids after surgery has been targeted as one method to reduce opioid diversion, prompting several groups to publish guidelines for surgical providers. Adoption and compliance with these guidelines will require buy-in from providers. However, little is known about the beliefs and practices of surgical providers regarding to patients’ pain, opioids, and prescribing patterns.

Methods:

We developed and distributed a 35-item survey to measure the beliefs, attitudes, and practice patterns of surgical providers regarding opioid prescribing. The survey was distributed to providers in five departments at a single tertiary academic medical center. Characteristics were generated. Additionally, we assessed awareness of and self-reported compliance with national standards for the prescription of opioids.

Results:

Preliminary results showed that the majority of respondents were male (56%). Seventy-eight (49%) respondents identified as attending physicians, while seven (5.1%) were fellows, twenty-five (18.2%) residents, 32 (23.4%) advanced practice clinicians, and 3 (2.2%) midwives. 56% of respondents reported being worried about patients having pain on nights and weekends when pharmacies were not open. 43% felt that it was difficult for patients to see providers for follow-up and opioid refills. 26% agreed that it is easier to give more opioids to minimize potential obstacles for post-operative pain control. A majority (53%) responded that they were more concerned about saving the patient's life than the possibility of opioid addiction. Although 44% felt that rural patients had trouble filling opioid prescriptions because of long travel distances, only 5% reported giving more pills to rural patients compared to urban patients. In contrast, 18% gave more pills to cancer patients than to non-cancer patients.

The majority of providers (51.1%) do not provide counseling regarding the use of opioids and only 20% provided their patients with a tapering schedule. 47% of providers said that they currently use the CDC guidelines for opioid prescribing in their practices, while 18% were unaware of the guidelines. 80% of providers responded that they do not have adequate time to counsel patients regarding opioid use and disposal. When asked who should be responsible for counseling patients about opioid disposal, 42% felt that pharmacists should be responsible for counseling. 67% of providers felt that patients were ultimately responsible for opioids and their proper disposal.

Conclusion:

Wide variation in the beliefs, attitudes, and practices exist among surgical providers. Despite previously reported evidence suggesting that surgical providers are over-prescribing opioids, providers often feel that they do not have the time nor hold primary responsibility for opioid stewardship. Further research is needed to develop and implement effective interventions for opioid stewardship after surgery.

52.17 Evaluation of Preventable Risk Factors for Trauma Mortality in Western Kenya

Posted on January 31, 2019

K. Carpenter1, C. H. Keung1,2, E. Rutto2, E. Chepkemoi2, J. Hogan3, H. W. Li1,4, J. Kisorio2  1Indiana University School Of Medicine,Department Of Surgery,Indianapolis, IN, USA 2Moi University,Department Of Surgery And Anaesthesiology,Eldoret, UASIN GISHU COUNTY, Kenya 3University of Alberta,Office Of International Surgery,Edmonton, ALBERTA, Canada 4Indiana University School Of Medicine,Indianapolis, IN, USA

Introduction:
Ninety percent of the 5 million annual deaths from traumatic injuries occur in low- and middle-income countries (LMICs). Kenya is one such nation, with limited trauma prevention policies and infrastructure. Road traffic accidents (RTAs), in particular, are a frequent cause of traumatic injury and subsequent mortality which continue to increase in incidence. This loss disproportionately impacts young people. Moi Teaching and Referral Hospital (MTRH) is a national tertiary referral hospital in Eldoret, Kenya, serving Western Kenya. In order to characterize the scope of traumatic injuries in this region, a prospective trauma registry was implemented at MTRH in September 2017. The purpose of this study was to identify how preventable risk factors affect mortality in RTAs in order to provide evidence-based recommendations for primary and secondary trauma prevention efforts.

Methods:
Data were collected prospectively on all patients presenting to MTRH primarily for acute traumatic injuries beginning in September 2017. Trauma registry data were retrospectively reviewed for patients involved in RTAs. Pedestrians struck were excluded. Data collected included mechanism of injury. Exposures of interest were seatbelt use, helmet use, and alcohol consumption. The outcome of interest was all-cause mortality within 30 days of presentation to MTRH. After patients with missing exposure data were excluded, Chi-square analysis and odds ratios were calculated using SAS 9.4.

Results:
Between September 2017 and April 2018, 1841 patients presented to MTRH following a traumatic injury. Of these, 870 patients were involved in an RTA. Exposure data was unknown or not recorded for 339 patients. 389 patients were involved in motorcycle accidents. Of these, only 10 patients (2.6%) were wearing a helmet at the time of the injury. No mortalities were recorded among helmeted patients. Unhelmeted patients had a 7.79% mortality rate and 2.5 greater odds of dying, though this was not significant (95% CI: 0.14-44.68, p=0.6002). 463 patients sustained injuries in RTAs involving automobiles. Only 2 (0.4%) were restrained and these both survived. Mortality among unrestrained patients was 7.1%. No significant relationship between seatbelt use and mortality was identified (p=0.2833). 17.2% of RTA injuries involved alcohol use. These patients had increased odds of mortality compared to those who had not consumed alcohol, which was not statistically significant (OR 1.41, 95% CI 0.65-3.09, p=0.3841). 

Conclusion:
Seatbelt and motorcycle helmet use is exceedingly rare among trauma patients in Western Kenya, while alcohol use is high. No mortalities were recorded among patients using a seatbelt or wearing a motorcycle helmet. Trauma prevention efforts in Western Kenya should target all three of these risk factors. More thorough patient data collection will be necessary to accurately monitor the success of such programs.

 

52.16 Nationwide trends in laparoscopic synchronous resection of colon cancer with liver metastasis

Posted on January 31, 2019

S. T. Lumpkin1, P. D. Strassle1,2, L. N. Purcell1, N. Lopez3, K. B. Stitzenberg1  1University Of North Carolina At Chapel Hill,Surgery,Chapel Hill, NC, USA 2University Of North Carolina At Chapel Hill,Epidemiology,Chapel Hill, NC, USA 3University Of California – San Diego,Surgery,San Diego, CA, USA

Introduction: Minimally invasive synchronous resection of primary colon cancer and metastatic liver lesions has been proven safe and effective. We hypothesized that the proportion of synchronous resections performed laparoscopically has increased.

Methods: Using the National Inpatient Sample (NIS) from 2009-2015, we identified all adult patients with colon cancer undergoing colon resection alone and those undergoing a synchronous colon and liver resection based on ICD-9 codes.  We compared the pace of laparoscopic uptake in the synchronous resection cohort to our control group, colon resections. All surgeries were classified as either laparoscopic or open. The yearly incidence of laparoscopic procedures was calculated using Poisson regression. Chi-square and Wilcoxon tests were used to compare patient and hospital characteristics.  

Results: Overall, 86,520 patients with colon cancer were identified, 55,766 underwent colon resections alone and 754 underwent synchronous resections. Wedge resections composed 50% of liver procedures. Laparoscopic procedures constituted 27,158 (49%) of the colon resections and 161 (21%) of synchronous resections, p<0.0001. Laparoscopic procedures have increased significantly in both colon resection (42% to 54%) and synchronous resections (11% to 32%) between 2009 and 2015, p<0.0001 and p=0.006, respectively, (Figure), although there was no significant difference in the pace of uptake of laparoscopy between groups (p=0.09).  Robotics composed 3% of all operations, and robot use was similar between colon resection and synchronous resections (p=0.13). Among synchronous resections, patients undergoing laparoscopic and open procedures were similar in regards to age (p=0.26), sex (p=0.69), race/ethnicity (p=0.28), insurance status (p=0.52), median household income (p=0.30), Charlson Comorbidity Index (p=0.19), and hospital size (p=0.95). However, significant differences were seen across colectomy procedure (p=0.004), liver procedure (p=0.0001), and hospital region (p=0.04). Specifically among synchronous resections, a laparoscopic approach was more likely performed in patients undergoing a left hemicolectomy (29% vs. 17%), liver ablation (32% vs. 18%), and among patients having surgery in the West (30% vs 19%). Laparoscopic approach was significantly less common among patients undergoing right hemicolectomy (19% vs. 25%, p=0.049). No difference was seen across teaching hospital status (21% vs. 24%, p=0.36).

Conclusions: Laparoscopic synchronous resection of colon and liver disease for colon cancer is becoming increasingly popular nationwide. The type of colon resection and liver procedure performed may guide a surgeon’s operative approach. There are also regional differences in practice patterns.

52.15 Family Can Hurt You the Most: Examining Perpetrators in Multiple Casualty Events

Posted on January 31, 2019

L. McLafferty1, J. Abolarin1, H. Carmichael1, C. G. Velopulos1  1University Of Colorado Denver,Department Of Surgery,Aurora, CO, USA

Introduction:

Recently, multiple casualty events, particularly mass shootings, have become a focus of media attention.  Because these events are heavily publicized, this can lead to the impression that school shootings and other rampages resulting in the deaths of strangers are typical. We hypothesized that many multiple homicides actually involve victims known to the perpetrator, such as family members or intimate partners.

Methods:
We analyzed homicides from the National Violent Death Reporting System (NVDRS) from 2003 to 2015.  We examined multiple homicide events (involving two or more victims), including those that were followed by suicide of the perpetrator or death by legal intervention (law enforcement). When the relationship between the perpetrator and victims was unknown, narratives were examined for evidence to classify incidents into categories including drug related, robbery or burglary related, argument or retaliation related, or mental health related incidents.

Results:
We identified 2425 incidents involving a total of 5,424 homicide victims, indicating that 9.3% of all homicide victims in NVDRS were killed in events that involved at least two victims.  Of these events, 13.8% (n=341) were homicides followed by suicide of the perpetrator.  Many of these incidents involved intimate partners or family members of the victims (n=741, 30.6%).  For those where the victims and the suspect were not family members (n=1684), homicides appeared to have been related to a combination of drug-related (n=247, 14.7%), robbery or burglary-related (n=289, 17.2%), argument or retaliation related (n=320, 19.0%).  Few of these events were related specifically to a mental health crisis (n=39, 2.3%) resulting in the deaths of multiple strangers. Only 31 incidents (1.3%) involved 5 or more victims. Of these, many still involved family members or intimate partners (n=14, 45%).  Of the mass casualty events involving strangers (n=, a larger portion were related to mental health (n=6, 35%).

Conclusion:
In our examination of multiple casualty events, we found that many of these still involve a single perpetrator killing an intimate partner and/or other family members, event for events with more than 4 victims. While the scenario of a single perpetrator with mental health issues going on a “shooting rampage” resulting in the deaths of multiple strangers is certainly more common in mass casualty incidents, it is rare overall.  Because this type of mass shooting event has recently been a focus of media attention, it is important to remember that these widely publicized incidents may not be typical of homicides involving multiple victims in the United States.

52.14 Risk Score for Post-Surgical Sustained Opioid Use

Posted on January 31, 2019

M. A. Chaudhary1, N. Bhulani1, D. Sturgeon1, N. K. Kwon1, E. D. Jager1, T. P. Koehlmoos2, A. J. Schoenfeld1, A. H. Haider1  1Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 2Uniformed Services University Of The Health Sciences,Bethesda, MD, USA

Introduction:
Post-surgical opioid prescription is widely maintained to be associated with the burgeoning opioid epidemic. Literature on appropriate post-surgical opioid prescription practices are now making way into policy. Stratification tools for identifying patients at risk of sustained opioid use are not available. In this context, we sought to generate a robust risk-score for identifying post-surgical patients susceptible to sustained opioid use.

Methods:
The Military Health System Data Repository (MDR) was queried for TRICARE claims data (2006-2014). Adult (18-64 years) patients who underwent one of 10 common general surgical, cardiovascular, orthopedic or urological procedures were included in the study. Opioid use 6 months prior to the index procedure and up to 6 months following discharge was assessed. A 75% random sample was generated from the study cohort for model generation and the remaining 25% sample was used for internal validation. Multivariable logistic regression models were utilized determine the predictors of sustained opioid use and a 100-point risk-score was generated using variables available at hospital discharge. C statistic and calibration curves were used to determine model performance.

Results:
A total of 86,356 patients records were included in the study. Among these, 64,767 randomly selected records were used for risk-score generation and 21,589 for internal validation. A total of 7.4% (n=6,365) of patients met criteria for sustained use after the index surgical procedure. The logistic regression model using only variables available at discharge provided robust identification of sustained post-surgical opioid use (area under the receiver operator curve=0.73). The risk-score included the following variables: age [{25-34, score (s):4}{35-44, s:6}{45-54, s:5}{55-64, s:4}], sex (female, s:4), race (non-white, s:-2), lower socio-economic status (s:7), prior opioid use (s:26), comorbid diagnosis [{diabetes, s:5}{depression, s:5}{anxiety, s:7}], procedure type [{urological, s:4}{orthopedic, s:7}], ICU admission (s:5), length of stay (>3 days, s:4) and discharge disposition (non-home discharge, s:13). The risk-score was further categorized based on the likelihood of sustained opioid use (Table). The risk-score performance in the validation sample was at par with the test sample.

Conclusion:
This study developed and validated a risk-score for sustained post-surgical opioid use that may have significant utility for clinical practice in discharge planning and engagement of pain management services for at-risk patients. Timely identification and appropriate outpatient care planning may reduce the incidence of sustained post-surgical opioid use in these patients.
 

52.13 An Uncontrolled Donation After Cardiac Death Program Improves Access to Kidney Transplantation

Posted on January 31, 2019

R. Choudhury1, H. B. Moore1, K. Prins1, T. Nydam1  1University Of Colorado Denver,Transplant Surgery,Aurora, CO, USA

Introduction:  Uncontrolled donation after cardiac death (uDCD) is a novel method to increase the supply of kidney allografts.  As opposed to controlled DCD (cDCD), uDCD remains an underutilized practice in the United States.  Its use in Europe, primarily Spain, has largely been restricted to out of hospital cardiac arrests (OHCA) with limited volumes of recoverably allografts.  Given the high amount of unsuccessful resuscitations following trauma arrests (URTA) in the United States, various groups have suggested that this population should be included in a uDCD program.  Estimates vary as to the number of potential kidney allografts which could be gained with this method.  Furthermore, the impact on the rate of transplant and overall survival for a patient currently on the kidney transplant waitlist are unknown.  The objective of this study was to estimate the impact of uDCD on the rate of transplant, chance of remaining on dialysis, and death for end stage renal disease (ESRD) patients on hemodialysis in the United States.  

Methods:  A decision analytic Markov state transition model was created using medical decision-making software (DATA 3.5, TreeAge Software, Inc., Williamstown, MA) in order to simulate three clinical scenarios for a group of 60,000 ESRD patients on hemodialysis who do not have access to a living donor (20,000 in each group).  Three clinical scenarios were modeled: 1) Reject uDCD: Patients are on kidney transplant waitlist and will never accept a uDCD kidney, 2) OHCA simulation: On waitlist and will accept a OHCA uDCD kidney if available, whose availability is estimated from the high volume uDCD European center (Spain), and 3) UTRA simulation: On waitlist and will accept a URTA uDCD kidney if available, whose local availability is estimated from the experience of high volume level one trauma center in the United States (Denver Health).  Markov model transition probability were calculated from the literature for “Reject uDCD” and OHCA simulations, and were derived from chart review of Denver Health Medical Center for the UTRA simulation.

Results: A UTRA uDCD program increases the rate of patients who are transplant at five years (24.3%, UTRA vs 21.3%, OHCA, vs 20.2%, Reject uDCD).  Furthermore, patients who remain on dialysis are also reduced in the UTRA simulation.  However, 5 year all-cause mortality is similar among groups (28.1%, 28.2%, 28.3%).

Conclusion: uDCD improves access to transplant for ESRD patients on the kidney transplant waitlist.  However, all-cause mortality is similar for patients who reject uDCD suggesting that careful patient selection is required to match a potential uDCD kidney allograft to patients who would likely not be offered a transplant by any other means such that net utility may be gained.
 

52.12 A Mobile Text App Increases Physician Participation in Safety Reporting in an Academic Medical Center

Posted on January 31, 2019

A. E. Graham1, I. Benjenk1, B. Umapathi1  1George Washington University School Of Medicine And Health Sciences,Washington, DC, USA

Introduction:
The landmark Institute of Medicine report “To Err is Human” was released in 1999 and since then increased attention has been placed on healthcare incident reporting. Despite concerted efforts of hospitals and other organizations, improvements in patient care derived from incident reporting have lagged behind goals. At most hospitals, the overwhelming majority of incident reports are placed by allied providers. However, physician participation is required to get a balanced view of system errors in large academic institutions.  Young physicians in particular need to be involved in reporting, evaluation, and system-based practice change. One of the largest hurdles to physician participation is access to reporting platforms. To address this, a new feature of the hospital’s HIPAA compliant texting application was added in the fall of 2017 so that providers could report incidents on their personal mobile phones instead of going through a computer-based application.

Methods:
The study was conducted at George Washington University Hospital. Monthly total incident reports from pre-intervention January 2017- June 2017 and post-intervention January-June 2018 were reviewed and categorized by method of reporting (computer vs. text) and position of reporter (physician, resident, other). An anonymous survey was given to 39 resident physicians to further evaluate resident participation. Descriptive statistics were performed in Excel and SPSS.

Results:
There was a 70.4% increase in incident reporting between our pre- and post-intervention periods (from 3090 to 5265). In the post-intervention period, 14.25% of incident reports were submitted through secure text message (750 reports). Of the 750, 585 (78%) were from physicians. Of the increase in total incident reporting, 34.4% was due to text message reporting and 26.8% was due to text message reporting by physicians. Of the residents surveyed, 53% reported submitting a report.  Of those who had reported an incident, 95.2% reported it as “very easy” or “fairly easy” to submit a report via text message, as compared to 9.5% who reported it as “very easy” or “fairly easy” to submit a report via computer application (p <0.01). Of those who had reported an incident, 76% reported having submitted more reports via text message than via computer, 14.2% reporting similar amounts, and 9.5% submitting more via computer.

Conclusion:
Institution of a mobile text app significantly increased physician participation in patient incident reporting. Our study shows that physicians, especially resident physicians, are more comfortable with mobile app-based patient safety reporting, and this is a viable model for the future.
 

52.11 Is Laparoscopic Adrenalectomy as Safe as Cholecystectomy? A propensity-matched analysis of NSQIP.

Posted on January 31, 2019

J. Limberg2, K. D. Gray1,2, T. Ullmann1,2, D. Stefanova1,2, B. M. Finnerty1,2, J. L. Buicko1,2, R. Zarnegar1,2, T. J. Fahey1,2, T. Beninato1,2  1New York Presbyterian Hospital,Surgery,New York, NY, USA 2Weill Cornell Medical College,Surgery,New York, NY, USA

Introduction:
Laparoscopic adrenalectomy (LA) is regarded as the treatment of choice for small, functioning tumors of the adrenal cortex.  Despite evidence that surgical cure of primary hyperaldosteronism is superior to medical treatment, there remains reluctance to refer patients for LA, with some recent reports advocating long-term medical management.   Here we aimed to determine whether the safety profile of LA is similar to that of elective laparoscopic cholecystectomy (LC), one of the most commonly performed laparoscopic procedures worldwide.

Methods:
Patients undergoing LA or LC between 2012 and 2015 were selected from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database. Patients who had a LA for an adrenal adenoma or aldosteronoma were compared to those who had LC for biliary colic. Propensity scores with nearest neighbor matching were calculated to determine propensity of receiving LA controlling for age, gender, race, Hispanic ethnicity, American Society of Anesthesiologists score (ASA) and 5-Factor Modified Frailty Index Score. The 30-day readmission rate, incidence of any post-operative complication, as well as the incidence of specific complications such as myocardial infarction, stroke, sepsis, surgical site infection, pneumonia, unplanned reintubation, renal failure, urinary tract infection, clostridium difficile infection and pulmonary embolism were compared between groups.

Results:
A total of 83,928 patients underwent either LA or LC during the study period. There were 16,480 patients who met the inclusion criteria and were successfully matched (n=1,490 LA, n=14,990 LC). The overall mortality rate was 0.02%, with no difference between groups (p=0.084). After propensity matching, the mean age of patients undergoing LA was 47.5 years compared with 50.5 years in the LC group, and they contained 80.9% and 76.9% female patients, respectively.  More LC patients were Hispanic (15.7% versus 12.9% LA). Patients in both groups had a similar rate of ASA score >3 (25.8% LA versus 24.7% LC) and Frailty Index Score >2 (7.0% LA versus 9.5% LC). There was no difference in the incidence of any post-operative complication between the two groups. The incidence of myocardial infarction (p=0.026) and wound infection (p<0.001 and p=0.002 for superficial and deep, respectively) were increased with LC. There were no statistically significant differences for the other specific post-operative complications between the two matched groups.

Conclusion:
In a propensity-matched comparison of patients from the ACS-NSQIP database undergoing LA and LC, there were no differences in the rate of overall postoperative complications and a slightly increased rate of wound complications and MI in patients undergoing LC.  Physicians should consider LA to have an equivalent risk profile to LC when deciding whether to refer patients to surgery.
 

52.10 Grading Complications of Operative and Non-operative Management of Acute Diverticulitis

Posted on January 31, 2019

A. A. Radwan1, S. Wei1, K. M. Mueck1, C. Wan1, C. E. Wade1, T. C. Ko1, S. G. Millas1, L. S. Kao1  1McGovern Medical School at UTHealth,Acute Care Surgery,Houston, TX, USA

Introduction:  There is no accepted grading system for comparing the risks between non-operative and operative management of surgical diseases. Recently, the Clavien-Dindo score for grading post-operative complications after elective surgery was modified and validated in hospitalized trauma patients who received non-operative and operative management. We hypothesized that this Adapted Clavien-Dindo in Trauma (ACDiT) scale can be used to grade and compare complications in acute diverticulitis patients across management strategies.

Methods:  We performed a retrospective cohort study of patients hospitalized for acute diverticulitis between 2011 – 2016 at a safety-net hospital. Baseline demographics and hospitalization data were collected. ACDiT grades were assigned to all patients. Grades ranged from 0 to 5b; a grade of 0 means no deviation from initial management plan, while a grade of 5b means hospital death despite active treatment. Univariate analysis was performed to compare baseline demographics in non-operatively versus operatively managed patients. ANOVA was used to determine differences in hospital-free days (HFD) based on ACDiT grades. HFD is defined as the number of days spent outside of any healthcare facility within 30 days of the initial admission. Linear regression was performed to assess correlation between ACDiT grades with HFD.

Results: There were a total of 260 patients, of which 177 (68%) were non-operatively and 83 (32%) were operatively managed. There were no differences in age, sex, race, Charleston Co-morbidity Index, or intraabdominal drain placement based on management strategy (p > 0.05). Eighty-five (33%) patients developed a complication, of which 78 (92%) complications received ACDiT grades 1 – 3b. There were no patients with ACDiT grade 5. Higher ACDiT grades correlated inversely with less HFD (r = -0.77, p < 0.0001), in the entire cohort (Figure) and individually for non-operative (r = -0.79, p <0.0001) and operative (r = -0.71, p <0.0001) management.

Conclusion: The ACDiT grading system can be used for classifying the severity of complications in hospitalized diverticulitis patients managed non-operatively and operatively and correlates with hospital-free days. ACDiT is a promising tool for weighing the risks of non-operative and operative management.
 

52.09 A Database of Trauma Funding Sources: Application for the Surgeon Health Policy Advocate

Posted on January 31, 2019

S. Lin1, C. Johnson3, A. E. Liepert2  1Oregon Health And Science University,Surgery,Portland, OR, USA 2University Of Wisconsin,Surgery,Madison, WI, USA 3American College Of Surgeons,Division Of Advocacy And Health Policy,Washington, D.C., USA

Introduction:  The implementation of trauma systems is a mainstay in American health care delivery resulting in many saved lives. It is regulated at the state level; however, a database of trauma system legislation and state funding sources has not been compiled in the medical literature. This absence reflects the disconnect physicians experience with legislative and regulatory processes, although they greatly impact the system of care they function within. This study attempts to compile trauma system regulations and funding sources in all 50 states to be used for Surgical Health Policy Advocacy in order to ensure adequate trauma care delivery and parity between and across states.

Methods:  Department of Health offices of each state’s trauma system were contacted via email and telephone. Resultant trauma system data were categorized according to the ability to access trauma system funding data, the presence of a trauma system in the state code, the presence of state statutes addressing trauma funding, the amount of trauma funding and the sources of funding for each trauma system. Data for each group included whether information could be compiled for an individual trauma system, the presence of state statutes addressing creation of a trauma system, the presence of statutes specific to trauma system funding and the total amount in dollars to trauma systems when available. States that underwent ACS trauma systems consultation were also captured.

Results: This taxonomy of trauma system funding was obtained for 24 (48%) states. This was fully compiled in 8 (16%) states, partially compiled in 16 (32%) states and no accurate data was available in 26 (52%) states. Specific legislation creating a trauma system was identified in 43 (86%) states.  No direct legislation of the formation of a trauma system was found in 5 (10%) states. 2 (4%) states do not have legislation creating a trauma system. 18 (36%) states have legislation directly funding their trauma systems, 4 (8%) states have legislation regarding trauma funding. No trauma system funding data was located in 28 (56%) states. Financial contributions ranged from $315,000 in Minnesota in 2011-2012 to $25,899,450 in Arkansas in 2016-2017. ACS Trauma Consultations took place at least once in 33 (66%) states. No correlation between states with these consultations and presence of trauma funding legislation was noted. 

Conclusion: This study demonstrates the difficulty of acquiring legislative information for use in Surgical Health Policy Advocacy efforts. It emphasizes the need for the development of tools to be used to summate and disseminate comprehensive and comparative legislative data and information. Future work to compile data such as this will aid in surgeon led legislative and regulatory advocacy efforts and improve not only systems as a whole but ultimately patient care.  

52.08 Communication Needs Among Staff Caring for Critically Injured: A Qualitative Study

Posted on January 31, 2019

A. M. Stey1,2, P. Liu1, C. Wybourn1,2, T. Bongiovanni1,2, R. Menza1,2, V. Singh1,2, T. Cage1, N. Brennan2, G. Ryan3  1University of California San Francisco,San Francisco, CA, USA 2Zuckerberg San Francisco General Hospital,San Francisco, CA, USA 3RAND Health,Santa Monica, CA, USA

Introduction:  The care of critically injured patients is complex and requires coordination across professionals from a wide range of disciplines including multidisciplinary staff. The aim of this study was to map out existing communication pathways and how they could be improved upon in the intensive care units.

Methods:  A total of 21 semi-structured case-based interviews were performed in an open mixed neuro and surgical intensive care unit in a level 1 trauma urban academic-affiliated safety net hospital. Neurosurgery, trauma surgery team members, neurologists, intensivists, nurses, pharmacists, and respiratory therapists of diverse backgrounds and seniority were interviewed. Interviewees were presented with 4 case scenarios involved decision making around two competing priorities that would require multi-disciplinary communication to determine treatment and timing of intervention and asked to describe what they would do.  The interviews responses were reviewed and range and central tendency are reported. 

Results: The central tendency theme of shared responsibility among all multi-disciplinary staff was identified. There were a range of roles each provider played in the care of critically injured patients. The primary team were the actualizers of decisions and interventions. In addition to the actualizers, two other roles were identified including secondary decision makers and interveners. Secondary decision makers were often consultants whose opinions were perceived as indispensable prior to decision making or intervention.  Interveners were frontline staff who were able to intervene if they felt a decision or intervention were discordant from clinical practice. There did seem to be considerable variation in the extent to which secondary decision makers and interveners could be engaged with the actualizers. A second important theme was the long communication routes among the teams. Typically, communication occurred between frontline staff between the teams and had to travel up and down hierarchies within the team prior to decision regarding treatment were implemented. When patients were gravely ill or disagreements/miscommunication through the standard pathway occurred, staff at the head of the hierarchies would reach out to their homologues directly to shorten the communication route. This seemed to resolve most disagreements due to a perception of mutual respect and perceived importance in the gesture between staff at the head of the hierarchies.

Conclusion: This study identified two themes including shared responsibility and long communication routes among teams. More research is needed to determine how communication could be improved upon both by promoting involvement of secondary decision makers and interveners as well as streamlining communication routes.

 

52.07 Outcomes following Major Oncologic Operations for non-AIDS Defining Cancers in the HIV Population

Posted on January 31, 2019

A. Chi1, B. E. Adams2, J. Sesti1, S. Paul1, A. Turner1, D. August1,3, D. Carpizo1,3, T. Kennedy1,3, M. Grandhi1,3, S. K. Libutti3, S. Geffner1, R. C. Langan1,3  1Saint Barnabas Medical Center,Surgery,Livingston, NJ, USA 2United States Military Academy,Mathematical Sciences,West Point, NEW YORK, USA 3Rutgers Cancer Institute Of New Jersey,Surgical Oncology,New Brunswick, NJ, USA

Introduction: Human immunodeficiency (HIV) patients are now living longer due to the availability of antiretroviral therapies and non-AIDS defining cancers are becoming more prevalent in this patient population.  However, a paucity of data remains on post-operative outcomes following resection of non-AIDS defining cancers in the HIV population.  

Methods: The National Inpatient Sample was utilized to identify patients who underwent gastrectomy, hepatectomy, pancreatectomy, colectomy or pulmonary resection for malignancy from 2005 to 2015 (HIV, N=52,742; non-HIV, N=11,885,184).  Complications were categorized by international classification of diseases (ICD)-9 diagnosis codes.  The HIV and non-HIV cohorts were matched on type of insurance, household income, zip code and urban/rural setting.  Logistic regression with the Survey Package in R was utilized to assess whether HIV was an independent predictor of post-operative complications. The analysis conducted took into account the sample trend weights and stratification to ensure inferences determined from the sample data are applicable to the population.

Results: Uncorrected data found HIV patients to have an increased rate of complications following colectomy, hepatic lobectomy, pulmonary resection (segmental and lobe), gastrectomy and distal pancreatectomy.  However, univariate and multivariate logistic regression (Table 1) found HIV to only be an independent predictor of complications following pulmonary lobectomy (p=0.011; OR 2.93, 95% CI 1.29-6.73).  There were no observed differences in post-operative mortality. 

Conclusion: Our findings highlight the relative safety of major cancer surgery in the HIV population.  HIV status should not exclusively be used to prohibit oncologic resections however; care providers need be cognizant of the potential increased risk of post-operative complications following pulmonary lobectomy.  Future analysis is planned to further characterize the complications and assess length of stay. These findings are an initial insight into quality of care and outcome metrics on HIV patients undergoing major cancer operations and serve as a platform to assess whether HIV impacts other aspects of the continuum of care.  

 

52.06 Patient-specific Postoperative Opioid Prescribing: A Pre-Post Analysis of an Educational Intervention

Posted on January 31, 2019

H. N. Overton1, V. Valero1, J. F. Griffin1, J. P. Taylor1, K. Giuliano1, A. B. Blair1, B. Moeckli1, R. B. Fransman1, A. Graham1, A. AwadElkarim1, R. Beckman1, S. He1, J. Liu1, S. DiBrito1, M. C. Bicket2, M. A. Makary1, E. R. Haut1, B. C. Sacks1  1Johns Hopkins University School Of Medicine,Surgery,Baltimore, MD, USA 2Johns Hopkins University School Of Medicine,Anesthesiology,Baltimore, MD, USA

Introduction:  Recent data show that 1 in 16 patients become chronic opioid users after undergoing a surgical procedure, and 45% of patients are over-prescribed opioids at the time of hospital discharge compared to their inpatient opioid use. The primary objective of this quality improvement project was to improve rates of appropriate opioid prescribing at discharge at an urban, academic hospital.

Methods:  We performed a retrospective cohort study with pre-post analysis after an educational intervention for surgical residents and discharge nurses. Patients who underwent a surgical procedure and required postoperative inpatient admission were included. An initial month long enrollment was followed by a resident-led educational intervention to individualize prescriptions based on the amount of opioids used in the 24-hours prior to discharge. Additional information on patient age, length of stay, admission status, acetaminophen and/or NSAID scheduled on day prior to discharge, and acetaminophen prescribed at discharge was collected. Patients with Methadone use during the hospitalization and/or at the time of discharge were excluded. The primary outcome was the difference between the total morphine milligram equivalents (MME) used in the 24 hours prior to discharge and the total daily MME prescribed at discharge. The mean differences were compared in the pre- and post-intervention groups. Secondary outcomes were the frequency of scheduled non-opioid pain medication during the inpatient admission and at the time of discharge.

Results: The prescribing patterns for 80 patients in the pre- and 69 patients in the post-intervention group (n=149) were reviewed. There were no significant differences between the pre- and post-intervention groups for any of the selected patient characteristics. The univariate model was determined to be most predictive by stepwise selection. The primary outcome of difference in MME was significantly different between the groups with the post-intervention group having 12.1 times less difference in MME than the pre-intervention group (95% CI: -20.9, -3.2; p=0.01) (Figure). Rates (mean (SD)) of prescribing non-opioid pain medication (Acetaminophen +/- NSAID) did not significantly differ between groups both during the inpatient admission (pre: 0.75 (0.49), post: 0.74 (0.44); p=0.88)) and at the time of discharge (pre: 0.63 (0.49), post: 0.70 (0.46); p=0.37).

Conclusion: An educational intervention for surgical residents and discharge nurses on how to customize the amount of opioids prescribed at the time of discharge resulted in significantly less over-prescription of opioids after surgical procedures. Future work should include optimization of non-narcotic pain medication usage in all settings of post-operative care.

« Previous Page
Next Page »