Faculty Bibliography

Introduction: The decision of admitting a stable patient to the ICU after major operation currently relies on clinical judgment and local hospital policies. We programmed an artificial intelligence (AI) to determine the appropriate level of care after major operation and compared its performance with clinician's judgement.
Method(s): ICU admission was deemed "appropriate" when at least 1 of 15 criteria (eg re-intubation, prolonged hypotension, new-onset arrhythmia) was observed. Using Institutional data (512 patients, 87 clinical variables), we programmed an AI to predict when ICU admission would have been appropriate. We prospectively evaluated whether surgeon, anesthesiologist, intensivist, or AI was the most accurate predictor in determining appropriateness of ICU admissions across 50 patients undergoing major surgery (general, vascular, urological). Accuracy of predictions was compared using receiver operating characteristic curve analysis.
Result(s): ICU care was appropriate (at least 1 of 15 objective criteria met) in 9 of 50 patients. Artificial intelligence correctly triaged to the appropriate level of care 82% of patients (surgeon 70%, anesthesiologist 58%, intensivist 64%). Receiver operating characteristic curve analysis revealed that AI's triage was the most accurate (area under the curve [AUC] 0.82), followed by anesthesiologist's (AUC 0.70), intensivist's (AUC 0.69), and surgeon's (AUC 0.60). Overall, clinicians leaned toward over-triaging patients to the ICU (Table).
Conclusion(s): Our study provides the first evidence that AI can have a role in supporting clinical decisions on postoperative triage. In the future, more sophisticated platforms can become integrated in daily clinical practice. [Figure presented]
Copyright

Introduction: Increasing evidence suggests that disparities in outcomes exist among patients with traumatic brain injury (TBI), but much less is known about such disparities in the elderly. The objective of this study was to determine if race and insurance status are associated with mortality among elderly patients with isolated moderate and severe TBI.
Method(s): A 4-year retrospective analysis of the Trauma Quality Improvement Program database (2013-2016) was performed to identify patients aged 60 and older with isolated moderate or severe TBI. Patients were stratified by race and insurance status comparing demographic characteristics and outcomes. A logistic regression analysis was performed to determine the relationship between race, insurance status, and mortality among elderly patients with isolated moderate and severe TBI.
Result(s): A total of 27,951 patients with isolated TBI were identified. Of those, 7.8% were black with 50.2% having insurance and 79.5% were white with 45.3% having insurance. The overall mortality rate was 9.22% with no significant differences in Head AIS. Black patients with insurance were significantly older (73 vs 63, p<0.001) and had more comorbidities (1 [0,2] vs 0 [0,1], p=0.002) when compared with black patients without insurance. With the exception of age, no significant differences were found among white patients. After adjusting for confounding variables, black race was independently associated with decreased mortality (AOR 0.69, 95% CI 0.5-0.96, p= 0.016).
Conclusion(s): Black race, independent of insurance, is associated with decreased mortality among older adults with isolated moderate and severe TBI. The role of race in affecting mortality following TBI warrants further investigation.
Copyright

Introduction: Ileal pouch-anal anastomosis (IPAA) failure is most commonly associated with pouch excision and permanent ileostomy. Motivated patients may choose to undergo a complex series of operations to attempt a re-do IPAA. We hypothesized that the outcomes of re-do IPAA are comparable to index IPAA.
Method(s): Patients undergoing index and redo IPAA at a specialized inflammatory bowel disease center between September 2016 and February 2019 were included. Operative and short-term outcomes were compared.
Result(s): There were 200 patients (index, n = 100 ; redo, n = 100). Eighty patients in the index and 70 in the redo group had their IPAA. Eight and 72 patients underwent 2 and 3-stage index IPAA surgery, respectively. Outside hospital system referrals were common for the cohort of redo operations (63% vs 29%, p < 0.001). Age, sex, BMI, American Society of Anesthesiologists (ASA) score, and primary diagnosis were comparable between the groups. Redo IPAA was associated with increased intraoperative bleeding (median, 200 vs 300 mL, p = 0.02), operative time (median, 223 vs 258 min, p = 0.001), and length of stay (median, 5 vs 6.5 days, p = 0.008). Thirty-day morbidity (41% vs 54%, p = 0.11), readmission (26% vs 24%, p = 0.78), and reoperation (1% vs 0, p = 1) were similar. One patient in the index group and 2 in the redo group had pouch failure with a median follow-up of 14 months.
Conclusion(s): In specialized high-volume centers, complex redo IPAA can safely be performed, with equivalent perioperative morbidity, when compared with index IPAA creation. Though redo IPAA was associated with greater blood loss, longer operative times, and length of stay, short-term morbidity was not significant.
Copyright

Introduction: Numerous barriers prevent surgical faculty from leveraging existing evaluation systems to assess resident operative skill. Training programs disproportionately rely on lengthy, end-of-rotation assessments which are inconsistently completed, lack specifics pertaining to operative technique, and are subject to recall bias. Existing solutions for procedure-based evaluations require trainees to proactively seek out faculty - an unreliable and socially-awkward process. We designed a novel tool to improve procedural assessment efficacy by leveraging the electronic health record (EHR) to trigger questionnaires and link assessments to specific patient encounters using a mobile application that overcomes existing obstacles to completing operative skill assessments.
Method(s): The real-time resident assessment (RETRA) tool was deployed across an academic surgical department in August 2018. The "out-of-operating room (OR)" EHR event triggered push notifications to faculty smartphones, launching a survey with questions from the American Board of Surgery Operative Performance Rating System. We compared assessment volume, compliance, and dashboard usability around the implementation of RETRA.
Result(s): Comparing 5-month intervals before and after RETRA deployment, we observed increased volume of procedural assessments and improved faculty compliance. Additionally, dashboard interfaces of cumulative RETRA data offered improved user experiences over previous systems, as measured by an industry-standard usability scale.
Conclusion(s): RETRA represents a novel approach to automating assessment of resident operative performance, overcomes existing barriers, and increases volume of technical performance data. While some features of RETRA overlap with previously described electronic assessment applications, we believe this represents a first-in-kind iteration that combines a mobile interface, automated EHR-mediated triggers, and preserved linkage between clinical data and resident assessments.
Copyright

BACKGROUND:Injury care involves the complex interaction of patient, physician, and environment that impacts patient complications, level of harm, and failure to rescue (FTR). FTR represents the likelihood of a hospital to be unable to rescue patients from death after in-hospital complications. OBJECTIVE:This study aimed to hypothesize that error type and number of errors contribute to increased level of harm and FTR. METHODS:Patient information was abstracted from weekly trauma performance improvement (PI) records (from January 1, 2016, to July 19, 2017), where trauma surgeons determined the level of harm and identified the factors associated with complications. Level of harm was determined by definitions set forth by the Agency for Healthcare Research and Quality. Logistic regression was used to determine the impact of individual factors on FTR and level of harm, controlling for age, gender, Charlson score, injury severity score (ISS), error (in diagnosis, technique, or judgment), delay (in diagnosis or intervention), and need for surgery. RESULTS:A total of 2216 trauma patients presented during the study period. Of 2216 patients, 224 (224/2216, 10.10 %) had complications reported at PI meetings; of these, 31 patients (31/224, 13.8 %) had FTR. PI patients were more likely to be older (mean age 51.3 years, SE 1.58, vs 46.5 years, SE 0.51; P=.008) and have higher ISS (median 22 vs 8; P<.001), compared with patients without complications. Physician-attributable errors (odds ratio [OR] 2.82; P=.001), most commonly errors in technique, and nature of injury (OR 1.91; P=.01) were associated with higher levels of harm, whereas delays in diagnosis or intervention were not. Each additional factor involved increased level of harm (OR 2.09; P<.001) and nearly doubled likelihood of FTR (OR 1.95; P=.01). CONCLUSIONS:Physician-attributable errors in diagnosis, technique, or judgment are more strongly correlated with harm than delays in diagnosis and intervention. Increasing number of errors identified in patient care correlates with an increasing level of harm and FTR.

BACKGROUND: Small bowel cancer (SBC) is a rare entity that can be associated with Crohn's Disease. The incidence of SBC in patients with CD is increased by 18.75-fold compared to normal population. The pathogenesis of SBC in the setting of CD is not fully understood, but the disease has a poor prognosis due to diagnostic challenges associated with the primary disease. The aim of this study is to present 2 cases treated within a year in a high volume IBD center and to conduct a systematic literature review of small bowel adenocarcinoma (SBA) associated with small bowel CD.
METHOD(S): Systematic literature review was done by using MEDLINE and EMBASE databases and data regarding demographics, presentation, diagnosis, treatment and survival were extracted. Articles that did not clearly state the location of the Crohn's disease and type of the cancer were excluded.
RESULT(S): We identified 216 patients diagnosed with small bowel adenocarcinoma in the setting of small bowel Crohn's disease from 117 studies. In this review obstruction was the most common initial symptom (n = 82, 59%; data are missing in 77 patients). Other common symptoms were abdominal pain (n = 12), anemia-bleeding (n = 11), diarrhea (n = 10), and fistulas (n = 5). there were 206 patients with one adenocarcinoma and 10 patients with 2 different adenocarcinomas. Among the patients with only one tumor, 154 patients (74.7%) were found to have ileal SBA. The median time to diagnosis of SBA from the diagnosis of CD was 18 months (1-300 months, data on 10 patients were missing). Out of 129, 64 patients (49.6%) were diagnosed with cancer after the surgical procedure whereas 46 patients (35.6%) were diagnosed intraoperatively and 15 (11.6%) were diagnosed preoperatively. Four patients (3.2%) were diagnosed at the autopsy without any surgical interventions and data were missing in 87 patients. 36.7% (18/49 patients) of the patients operated for obstruction were alive at one year, and 15.2% (7/46 patients) at 2-year. Although these percentages are lower than other studies reported in the literature, data was missing in 34 patients, so this might have affected the outcomes. CONCLUSION(S): Small bowel adenocarcinoma should be in the differential diagnosis in patients with longstanding ileal Crohn's disease presenting with small bowel obstruction, anemia, and perforation. Diagnosis and management of the small bowel adenocarcinoma in the setting of Crohn's disease is challenging and awareness and early diagnosis may avoid mortality

BACKGROUND:Prolonged operating room turnover time erodes patient and employee satisfaction and value. METHODS:Lean and value stream mapping was applied to three operating room teams at an academic health center in New York City and a solution called Performance Improvement Team (PIT Crew) was piloted. RESULTS:Overall, 10% of operating room turnover steps were considered non-valued and were eliminated and 25% of previously sequential steps were performed synchronously. Seven institutional dogmas were eliminated, and three hospital policies were changed. After 35 pilot turnovers, median operating room turnover time improved from 37 minutes (range 26-167) in historical matched controls to 14 minutes (range 10-45, p<0.0001) for the PIT Crew. Cost of the PIT Crew was $1,298 daily and estimated return on investment was $19,500 per day. CONCLUSIONS:Lean and value stream mapping identifies non-valued steps in operating room turnover and affords opportunities for efficiency. Once institutional rules and dogma are changed, culture and workflow improve and turnover time significantly improves. This process adds cost but is profitable. Scalability and sustainability is under further study, as is the "halo effect" on the culture in other non-PIT Crew operating rooms.

INTRODUCTION/BACKGROUND:Innovative strategies to reduce costs while maintaining patient satisfaction and improving delivery of care are greatly needed in the setting of rapidly rising health care expenditure. Intensive care units (ICUs) represent a significant proportion of health care costs due to their high resources utilization. Currently, the decision to admit a patient to the ICU lacks standardization because of the lack of evidence-based admission criteria. The objective of our research is to develop a prediction model that can help the physician in the clinical decision-making of postoperative triage. MATERIALS AND METHODS/METHODS:Our group identified a list of index events that commonly grants admission to the ICU independently of the hospital system. We analyzed correlation among 200 quantitative and semiquantitative variables for each patient in the study using a decision tree modeling (DTM). In addition, we validated the DTM against explanatory models, such as bivariate analysis, multiple logistic regression, and least absolute shrinkage and selection operator. RESULTS:Unlike explanatory modeling, DTM has several unique strengths: tree models are easy to interpret, the analysis can examine hundreds of variables at once, and offer insight into variable relative importance. In a retrospective analysis, we found that DTM was more accurate at predicting need for intensive care compared with current clinical practice. DISCUSSION/CONCLUSIONS:DTM and predictive modeling may enhance postoperative triage decision-making. Future areas of research include larger retrospective analyses and prospective observational studies that can lead to an improved clinical practice and better resources utilization.

OBJECTIVE:The purpose of this study was to assess the use of apparent diffusion coefficient (ADC) during DWI for predicting complete pathologic response of rectal cancer after neoadjuvant therapy. MATERIALS AND METHODS/METHODS:A systematic review of available literature was conducted to retrieve studies focused on the identification of complete pathologic response of locally advanced rectal cancer after neoadjuvant chemoradiation, through the assessment of ADC evaluated before, after, or both before and after treatment, as well as in terms of the difference between pretreatment and posttreatment ADC. Pooled mean pretreatment ADC, posttreatment ADC, and Δ-ADC (calculated as posttreatment ADC minus pretreatment ADC divided by pretreatment ADC and multiplied by 100) in complete responders versus incomplete responders were calculated. For each parameter, we also pooled sensitivity and specificity and calculated the area under the summary ROC curve. RESULTS:/s, in complete and incomplete responders, respectively (p = 0.00001). The Δ-ADC percentages were also significantly higher in complete responders than in incomplete responders (59.7% vs 29.7%, respectively, p = 0.016). Pooled sensitivity, specificity, and AUC were 0.743, 0.755, and 0.841 for pretreatment ADC; 0.800, 0.737, and 0.782 for posttreatment ADC; and 0.832, 0.806, and 0.895 for Δ-ADC. CONCLUSION/CONCLUSIONS:Use of ADC during DWI is a promising technique for assessment of results of neoadjuvant treatment of rectal cancer.

Background/UNASSIGNED:Half of patients with pancreatic adenocarcinoma (PC) present with regionally advanced disease. This includes borderline resectable and locally advanced unresectable tumors as defined by current NCCN guidelines for resectability. Chemoradiation (CH-RT) is used in this setting in attempt to control local disease, and possibly downstage to resectable disease. We report a phase I/II trial of a combination of 5FU/Oxaliplatin with concurrent radiation in patients presenting with borderline resectable and locally advanced unresectable pancreatic cancer. Methods/UNASSIGNED:. Concurrent radiation therapy consisted of 4,500 cGy in 25 fractions (180 cGy/fx/d) followed by a comedown to the tumor and margins for an additional 540 cGy ×3 (total dose 5,040 cGy in 28 fractions). Following completion of CH-RT, patients deemed resectable underwent surgery; those who remained unresectable for cure but did not progress (SD, stable disease) received mFOLFOX6 ×6 cycles. Survival was calculated using Kaplan-Meier analysis. End-points of the phase II portion were resectability and overall survival. Results/UNASSIGNED:) was well tolerated and it was used as the recommended phase II dose. An additional 7 patients were treated in the phase II portion, 5 of whom completed CH-RT; the remaining 2 patients did not complete treatment because of grade 3 toxicities. Overall, 4/24 did not complete CH-RT. Grade 4 toxicities related to initial CH-RT were observed during phase I (n=2, pulmonary embolism and lymphopenia) and phase II (n=3, fatigue, leukopenia and thrombocytopenia). Following restaging after completion of CH-RT, 4 patients had progressed (PD); 9 patients had SD and received additional chemotherapy with mFOLFOX6 (one of them had a dramatic response after two cycles and underwent curative resection); the remaining 7 patients (29.2%) were noted to have a response and were explored: 2 had PD, 4 had SD, still unresectable, and 1 patient was resected for cure with negative margins. Overall 2 patients (8.3%) in the study received curative resection following neoadjuvant therapy. Median overall survival for the entire study population was 11.4 months. Overall survival for the two resected patients was 41.7 and 21.6 months. Conclusions/UNASSIGNED:Combined modality treatment for borderline resectable and locally advanced unresectable pancreatic cancer with oxaliplatin, 5FU and radiation was reasonably well tolerated. The majority of patients remained unresectable. Survival data with this regimen were comparable to others for locally advanced pancreas cancer, suggesting the need for more novel approaches.