Use of a national registry to define a composite quality metric for rectal cancer.

BACKGROUND: Quality assessment in oncologic surgery has traditionally involved reporting discrete metrics that may be difficult for patients and referring providers to interpret. We define a composite quality metric (CQM) for resection in rectal cancer. METHODS: We queried the National Cancer Database to identify patients undergoing low anterior resection for clinical stage II-III rectal adenocarcinoma between 2010 and 2017. CQM was defined as appropriate neoadjuvant therapy, margin-negative resection, appropriate lymph node assessment, postoperative length of stay (LOS) < 75th percentile, and no 30-day readmission or mortality. RESULTS: 19,721 patients met inclusion criteria; 8,083 (41%) had a CQM. The most common reasons for failure to achieve CQM: inadequate node assessment (27%), prolonged LOS (26%). On Cox modeling, CQM (aHR 0.70, 95% CI [0.66, 0.75]) was associated with improved overall survival. CONCLUSION: CQM is independently associated with improved survival in rectal cancer and may be an effective measure of quality.

The benefit of liver transplant beyond Milan criteria: An analysis using the National Cancer Database.

BACKGROUND: Published studies examining the efficacy of liver transplantation in patients presenting with hepatocellular cancer beyond the traditional Milan criteria for liver transplantation have primarily been single institution series with limited ability to compare outcomes to alternative methods of management. METHODS: We queried the National Cancer Database to identify patients presenting between 2004 and 2016 with histologically confirmed clinical stage III and IVA hepatocellular cancer. Multivariable regression was used to identify factors associated with liver transplantation. Patients undergoing liver transplantation were 1:1 propensity score-matched for age, demographics, comorbid disease, clinical stage, and histologic resection margin to those undergoing surgical resection. The Kaplan-Meier method was used to compare overall survival profiles for matched cohorts. RESULTS: Seven hundred and ninety-two patients met inclusion criteria-590 (74.5%) underwent surgical resection and 202 (25.5%) liver transplantation. On adjusted analysis, patients undergoing liver transplantation were less likely to have advanced age (>60 years; odds ratio 0.39, 95% confidence interval [0.21-0.71]) and to be of Black race (odds ratio 0.42, 95% confidence interval [0.23-0.73]) or Asian (odds ratio 0.25, 95% confidence interval [0.11-0.53]) ethnicity but were more likely to have advanced (Charlson score >2) comorbidity scores, (odds ratio 2.48, 95% confidence interval [1.58-3.90]) and more likely to have private health insurance (odds ratio 4.17, 95% confidence interval [1.31-18.66]) than those undergoing surgical resection. On Kaplan-Meier analysis of matched cohorts, patients undergoing liver transplantation demonstrated significantly better rates of 5-year overall survival (65.3% vs 26.3%, P < .0001) and longer median overall survival times than those undergoing resection (53.1 ± 2.78 vs 26.9 ± 1.20 months, P < .0001). CONCLUSION: Liver transplantation offers the potential to be an effective treatment modality in select patients presenting with stage III and IVA hepatocellular cancer.

Are Residents Able to Perform a Laparoscopic Cholecystectomy with Acceptable Outcomes When the Attending Is Not Scrubbed?

BACKGROUND: Recent socioeconomic pressures in healthcare and work hour resections have limited opportunities for resident autonomy and independent decision-making. We sought to evaluate whether contemporary senior residents are being given the opportunity to operate independently and whether patient outcomes are affected when the attending is not directly involved in an operation. STUDY DESIGN: The VA Surgical Quality Improvement Program (VASQIP) Database was queried to identify patients undergoing elective laparoscopic cholecystectomy between 2004 and 2019. Cases were categorized as "attending" or "resident" depending on whether the attending surgeon was scrubbed. Cohorts were 1:1 propensity score-matched (PSM) for demographics, comorbidities, and facility case-mix. Clinical outcomes for matched cohorts were compared by standard methods. RESULTS: There were 23,831 records for patients who underwent laparoscopic cholecystectomy; 20,568 (86%) performed with the attending scrubbed, and 3,263 (14%) without the attending scrubbed. Over time there was a significant decrease in the proportion of cases without the attending scrubbed, 18% in 2004-2009 to 13% in 2015-2019 (p < 0.001). On PSM, 3,263 patients undergoing laparoscopic cholecystectomy by the residents without the attending scrubbed were successfully matched (1:1) to cases with the attending scrubbed. On comparison of matched cohorts, procedures performed without the attending scrubbed were statistically longer (102 vs 98 minutes, p = 0.001) but with no difference in rates of postoperative complications (5% vs 5%, p = 0.9). CONCLUSION: In comparison with cases done with more direct attending involvement, residents perform laparoscopic cholecystectomies efficiently without increased complications. Over time, attendings are more frequently scrubbed for the operation.

Variation in Hospital Cost and 1-Year Episodes of Care after Diaphragmatic Hernia Repair.

BACKGROUND: Diaphragmatic hernia repair is a common operation performed at all types of hospitals. The variation in costs and repeat episodes of care after this operation is not known. STUDY DESIGN: The Healthcare Cost and Utilization Project State Inpatient and State Ambulatory Surgery and Services Databases for Florida were queried to identify patients undergoing diaphragmatic hernia repair between 2011 and 2018 and the associated inpatient and outpatient encounters within 12 months postoperatively. Hospitals were ranked by cost and grouped into quintiles. All costs and charges were reliability and case-mix adjusted with the use of hierarchical multivariable regression. RESULTS: In total, 8,848 patients underwent diaphragmatic hernia operations at 158 hospitals. The most expensive hospital quintile had lower surgical volume, location in rural settings, and fewer than 100 beds. There was a wide variation in costs after diaphragmatic hernia repair. On unadjusted comparison, index costs were $23,041 more expensive in hospitals in the highest quintile than in the lowest quintile. Cost differences were persistent even after case-mix and reliability adjustment. The variation in adjusted aggregate charges for associated outpatient and inpatient encounters in the first year after the index operation was considerably lower than that of the index hospitalization. CONCLUSION: There is nearly a 2-fold variation in the cost of a diaphragmatic hernia repair across hospitals. Most of the variation occurs during the index surgical encounter and not for repeat encounters during the first postoperative year. As bundled payment models mature, hospitals and payers will need to target this variation to ensure cost-efficiency.

Contemporary Matched-Cohort Comparison of Surgical Approach to Inguinal Hernia Repair: Are Minimally Invasive Approaches Associated with Higher Rates of Recurrence?

BACKGROUND: Current studies evaluating outcomes for open, laparoscopic, and robotic inguinal hernia repair, in general, include small numbers of robotic cases and are not powered to allow a direct comparison of the 3 approaches to repair. STUDY DESIGN: We queried the Veterans Affairs Surgical Quality Improvement Program Database to identify patients undergoing initial elective inguinal hernia repair between 2013 and 2017. Propensity score matching and multivariable logistic regression were used to make risk-adjusted assessments of association between surgical approach and outcome. RESULTS: A total of 39,358 patients underwent initial elective inguinal hernia repair; 32,881 (84%) underwent an open approach, 6,135 (16%) underwent a laparoscopic approach, and 342 (1%) underwent a robotic-assisted approach. Two hundred sixty-six (1%) patients had a recurrent repair performed during follow-up. On univariate comparison, patients undergoing a robotic-assisted approach had longer operative times for unilateral repair than those undergoing either an open or laparoscopic (73 ± 31 vs 74 ± 29 vs 107 ± 41 minutes; p < 0.001) approach. On multivariable logistic regression, patients with a higher BMI had an increased adjusted risk of a postoperative complication, but there was no association between surgical approach and complication rate. Three hundred forty-two patients undergoing robotic repair were 1:3:3 propensity score matched to 1,026 patients undergoing laparoscopic and 1,026 undergoing open repair. On comparison of matched cohorts, there were no statistical differences between approaches regarding recurrence (0.6% vs 0.8% vs 0.6%, p > 0.05) or complication rate (0.6% vs 1.2% vs 1.2%, p > 0.05). CONCLUSIONS: In patients undergoing initial elective inguinal hernia repair, rates of hernia recurrence are low independent of surgical approach. Both robotic and laparoscopic approaches demonstrate rates of early postoperative morbidity and recurrence similar to those for the open approach. The robotic approach is associated with longer operative time than either laparoscopic or open repair.

Outcomes of Transcatheter and Surgical Aortic Valve Replacement in Distressed Socioeconomic Communities.

Objective Patients of low socioeconomic status have an increased risk of complications following cardiac surgery. We aimed to identify disparities in patients undergoing aortic valve replacement using the Distressed Communities Index (DCI), a comparative measure of community well-being. The DCI incorporates seven distinct socioeconomic indicators into a single composite score to depict the economic well-being of a community. Methods The Healthcare Cost and Utilization Project State Inpatient Database (HCUP-SID) for Florida and Washington was queried to identify patients undergoing surgical and transcatheter aortic valve replacement (surgical aortic valve replacement [SAVR], transcatheter aortic valve replacement [TAVR]) between 2012-2015. Patients undergoing TAVR and SAVR were propensity-matched and stratified based on the quintile of DCI score. A distressed community was defined as those in quintiles 4 and 5 (at-risk and distressed, respectively); a non-distressed community was defined as those in quintiles 1 and 2 (prosperous and comfortable, respectively). Outcomes following aortic valve replacement were compared across groups in distressed communities. Propensity score matching was used to balance baseline covariates between groups. Results A total of 27,591 patients underwent aortic valve replacement. After propensity matching, 5,331 patients were identified in each TAVR and SAVR group. Distressed TAVR patients had lower rates of postoperative pneumonia (7.6% vs. 3.8%, p<0.001), sepsis (3.6% vs. 1.9%, p<0.05), and cardiac complications (15.4% vs. 7.5%, p<0.001) when compared to highly distressed SAVR patients. When comparing distressed SAVR and TAVR and low distressed SAVR and TAVR groups, no significant difference was found in postoperative outcomes, except distressed TAVR experienced more cases of UTI. Conclusions Highly distressed TAVR patients had lower incidences of postoperative sepsis, pneumonia, and cardiac complications when compared to the highly distressed SAVR cohort. Patients undergoing TAVR in highly distressed communities had an increased incidence of postoperative urinary tract infection. DCI may be a useful adjunct to current risk scoring systems.

Low and moderate grade retroperitoneal liposarcoma: Is adjuvant radiotherapy associated with improved survival in patients undergoing R1 resection?

BACKGROUND: Few studies evaluate the efficacy of adjuvant radiotherapy (aXRT) in patients with retroperitoneal liposarcoma undergoing resection to histologically positive (R1) margins. METHODS: We queried the National Cancer Database to identify patients undergoing R1 resection for localized, large (>5 cm) low and moderate grade retroperitoneal liposarcoma between 2004 and 2016. Kaplan Meier method was used to compare overall survival (OS) for patients receiving aXRT to a 1:2 propensity-matched cohort of patients undergoing resection alone. RESULTS: A total of 322 (76.5%) patients underwent R1 resection alone, while 99 (23.5%) underwent resection followed by aXRT. The 99 receiving aXRT were successfully 1:2 propensity-score matched to 198 undergoing resection alone. There was no difference in 5-year OS between matched cohorts (69.7% vs 76.2%, p = 0.40). CONCLUSIONS: In patients undergoing R1 resection of moderate- and well-differentiated retroperitoneal liposarcoma, use of aXRT is not associated with an improvement in OS.

Textbook oncologic outcome in pancreaticoduodenectomy: Do regionalization efforts make sense?

BACKGROUND AND OBJECTIVES: Few empiric studies evaluate the effects of regionalization on pancreatic cancer care. METHODS: We queried the National Cancer Database to identify patients undergoing pancreaticoduodenectomy for clinical stage I/II pancreatic adenocarcinoma between 2006 and 2015. Facilities were categorized by annual pancreatectomy volume. Textbook oncologic outcome was defined as a margin negative resection, appropriate lymph node assessment, no prolonged hospitalization, no 30-day readmission, no 90-day mortality, and timely receipt of adjuvant chemotherapy. Multivariable regression adjusted for comorbid disease, pathologic stage, and facility characteristics was used to evaluate the relationship between facility volume and textbook outcome. RESULTS: Sixteen thousand six hundred and two patients underwent pancreaticoduodenectomy; 3566 (21.5%) had a textbook outcome. Operations performed at high volume centers increased each year (45.8% in 2006 to 64.2% in 2015, p < 0.001) as did textbook outcome rates (14.3%-26.2%, p < 0.001). Surgical volume was associated with textbook outcome. High volume centers demonstrated higher unadjusted rates of textbook outcome (25.4% vs. 11.8% p < 0.01) and increased adjusted odds of textbook outcome relative to low volume centers (odds ratio: 2.39, [2.02, 2.85], p < 0.001). Textbook outcome was associated with improved overall survival independent of volume. CONCLUSIONS: Regionalization of care for pancreaticoduodenectomy to high volume centers is ongoing and is associated with improved quality of care.

Assessment of Textbook Outcome After Surgery for Stage I/II Non-small Cell Lung Cancer.

Outcomes after cancer resection are traditionally measured individually. Composite metrics, or textbook outcomes, bundle outcomes into a single value to facilitate assessments of quality. We propose a composite outcome for non-small cell lung cancer resections, examine factors associated with the outcome, and evaluate its effect on overall survival. We queried the National Cancer Database for patients with stage I/II non-small cell lung cancer who underwent sublobar resection, lobectomy, or pneumonectomy from 2010 to 2016. We defined the metric as margin-negative resection, sampling of ≥10 lymph nodes, length of stay <75th percentile, no 30-day mortality, no readmission, and receipt of indicated adjuvant therapy. Multivariable logistic regression, Cox proportional hazards modeling, survival analyses, and propensity score matching were used to identify factors associated with the outcome and overall survival. Of 88,208 patients, 70,149 underwent lobectomy, 14,922 underwent sublobar resection, and 3,137 underwent pneumonectomy. Textbook outcome was achieved in 26.3% of patients. Failure to achieve the outcome was most commonly driven by inadequate nodal assessment. Textbook outcome was more likely after minimally invasive surgical approaches (aOR = 1.47; P< 0.001) relative to open resection and less likely after sublobar resection (aOR = 0.20; P< 0.001) relative to lobectomy. Achievement of textbook outcome was associated with an 9.6% increase in 5-year survival (P< 0.001), was independently associated with improved survival (aHR = 0.72; P < 0.001), and remained strongly associated with survival independent of resection extent after propensity matching. One in 4 patients undergoing non-small cell lung cancer resection achieve textbook outcome. Textbook outcome is associated with improved survival and has value as a quality metric.

Supervised independence in elective inguinal hernia repairs performed by surgical residents is not associated with compromised short clinical outcome or rates of reoperation for recurrence.

BACKGROUND: We evaluate the association between attending surgeon involvement and clinical outcome in elective inguinal hernia repairs performed by residents. METHODS: Patients undergoing initial elective unilateral inguinal hernia repair between 2004 and 2019 were identified using the Veterans Administration Surgical Quality Improvement Program Database. The level of attending surgeon involvement was categorized as active (attending scrubbed [AS]) or passive (supervising the resident's performance but not scrubbed [ANS]). AS and ANS herniorrhaphies were 1:1 propensity matched for patient demographics, comorbidities, surgical approach, resident postgraduate level, and year of repair. Rates of complication and recurrence for matched cohorts were compared by standard methods. RESULTS: 30,784 patients met inclusion criteria. 5136 (17%) repairs were performed without the attending scrubbed. On comparison of matched-cohorts, overall complication rates (1.7% vs 1.2%, p = 0.07) and rates of recurrence (1.9% vs 1.4%, p = 0.041) for patients undergoing herniorrhaphy AS were statistically similar to those performed ANS. CONCLUSION: Supervised independence in elective inguinal hernia repair performed by surgical residents is not associated with inferior clinical outcomes.

Does resident autonomy in colectomy procedures result in inferior clinical outcomes?

BACKGROUND: The amount of time surgical trainees spend operating independently has been reduced by work-hour restrictions and shifts in the health care environment that impede autonomy. Few studies evaluate the association between clinical outcome and resident autonomy. METHODS: The Veterans Affairs Surgical Quality Improvement Program database was queried to identify patients undergoing partial colectomy for neoplasm between 2004 and 2019. Rectal resections, emergency procedures, and those involving postgraduate year 1 and 2 residents were excluded. Records were categorized as performed with the attending scrubbed or not scrubbed. Hierarchical logistic regression was used to identify factors independently associated with operative time, morbidity, and mortality. RESULTS: In total, 7,347 patients met inclusion criteria; 6,890 (93.6%) were categorized as attending scrubbed and 457 (6.4%) as attending not scrubbed. The cohorts were similar in terms of patient demographics, including age, race, body mass index, and American Society of Anesthesiologists class. There were no differences between cohorts in terms of operative time (attending not scrubbed 3.02 hours, attending scrubbed 3.07 hours, P = .42). On hierarchical logistic regression adjusted for age, gender, race, body mass index, functional status, cancer location, facility operative level, wound class, American Society of Anesthesiologists class, length of operation, operative modality (open or minimally invasive), postgraduate year of resident, and year, there were no differences in odds of complications, major morbidity, or mortality based on attending involvement. CONCLUSION: Colectomies performed by residents with appropriate levels of autonomy are efficient and safe. Our results indicate that attending surgeon judgment regarding resident autonomy is sound and that educational environments can be designed to foster resident independence and preserve clinical quality, safety, and efficiency.

Intrahepatic cholangiocarcinoma: Is there a role for liver transplantation?

BACKGROUND: Liver transplantation offers a potential for curative-intent treatment in patients presenting with non-metastatic intrahepatic cholangiocarcinoma that is not amenable to partial hepatectomy. There is little empiric evidence evaluating the efficacy of liver transplantation in patients with intrahepatic cholangiocarcinoma. METHODS: We queried the National Cancer Database to identify patients presenting with histologically confirmed clinical stage I to III intrahepatic cholangiocarcinoma between 2004 and 2016. Propensity scoring was used to develop matched cohorts of patients undergoing treatment with liver transplantation, surgical resection, or chemotherapy alone. Kaplan Meier methods were used to compare rates of overall survival. RESULTS: One thousand four hundred and eleven patients met inclusion criteria. Of these, 66 (4.7%) underwent liver transplantation, 461 (32.7%) underwent surgical resection, and 884 (62.6%) were treated with chemotherapy alone. On adjusted analysis, patients undergoing liver transplantation were more likely to be male (odds ratio 4.35, 95% confidence interval [0.12, 0.42]), have a Charlson Comorbidity Score ≥2 (odds ratio 3.11, 95% confidence interval [1.44, 6.57]), and to receive both neoadjuvant (odds ratio 2.78, 95% confidence interval [1.36,5.75], and adjuvant (odds ratio 1.94, 95% confidence interval [0.97, 3.87]) systemic therapy than those undergoing resection. On Kaplan Meier analysis, patients undergoing liver transplantation demonstrated rates of 5-year overall survival (36.1% vs 34.7%, P = .53) that were statistically identical to those for stage-matched and margin-matched patients undergoing resection but significantly better than those for stage-matched patients treated with systemic therapy alone (36.1% vs 5.3%, P < .0001). CONCLUSION: Patients undergoing liver transplantation for intrahepatic cholangiocarcinoma demonstrate overall survival profiles similar to stage-matched and margin-matched patients undergoing surgical resection. Liver transplantation is an effective treatment modality in select patients presenting with localized intrahepatic cholangiocarcinoma.

Small-sized gastrointestinal stromal tumors: Is there an association between expectant management and overall survival?

BACKGROUND: Small-sized gastrointestinal stromal tumors (GISTs) have limited malignant potential. Few studies evaluate the safety and efficacy of expectant management (EM) for patients presenting with small GIST. METHODS: We queried the National Cancer Database to identify patients ≤65 years presenting with GISTs smaller than 3 cm in size between 2004 and 2015. Patients undergoing EM were 1:3 propensity score matched for relevant covariates to patients undergoing resection. Kaplan-Meier (KM) analysis of matched cohorts was used to evaluate the association between EM and overall survival (OS). RESULTS: 1330 patients met inclusion criteria; 966 (72.6%) had gastric GISTs. 1196 (89.9%) underwent resection; 134 (10.1%) EM. 117 patients undergoing EM were propensity-matched to 356 patients undergoing resection. There was no difference in 5-year OS between patients undergoing EM and those undergoing resection (95.7% vs 92.6%, p = 0.4882). CONCLUSIONS: Survival for small GISTs is similar with expectant management or resection.

Resectable intrahepatic and hilar cholangiocarcinoma: Is margin status associated with survival?

BACKGROUND: Prior studies evaluating the effect of margin status on clinical outcome in patients undergoing resection for intrahepatic and extrahepatic hilar cholangiocarcinoma include small numbers of patients with histologically positive margins. The value of margin negative resection in these cases remains unclear. METHODS: We queried the National Cancer Database to identify patients undergoing resection for clinical stage I to III intrahepatic and extrahepatic hilar between 2004 and 2015. Patients receiving neoadjuvant therapy and those having <3 lymph nodes examined were excluded. Patients undergoing positive resection were 1:1 propensity matched to those undergoing negative resection. Kaplan-Meier methods were used to compare overall survival for the matched cohorts. RESULTS: In the study, 3,618 patients met the inclusion criteria, and 3,018 (83.4%) underwent negative resection; 600 (16.6%) positive resection. Patients undergoing negative resection had smaller tumors (2.97 ± 0.07 cm vs 3.49 ± 0.15 cm), were less likely to have stage 3 disease (16.7% vs 25.7%) and to receive adjuvant radiation (27.1% vs 45.7%) and chemotherapy (49.4% vs 61.0%) than those undergoing positive resection (all P < .05). On comparison of matched cohorts, patients undergoing negative resection had longer median overall survival (24.5 ± 0.02 vs 19.1 ± 0.02 months) and higher rates of 5-year overall survival (24.5% vs 16.7%) than those undergoing positive resection (P < .01). CONCLUSION: In patients presenting with resectable intrahepatic and extrahepatic hilar, negative resection is associated with improved overall survival. Extended resections performed in an effort to clear surgical margins are warranted in patients fit for such procedures.

Is positive histologic surgical margin associated with overall survival in patients with resectable gallbladder cancer?

BACKGROUND: Achieving microscopically negative (R0) surgical margins in gallbladder cancer often requires a partial hepatectomy with associated risk of morbidity and potential to delay adjuvant therapy. Prior studies on the importance of margin status in resectable gall bladder cancer include small numbers of patients with positive (R1) resection margins and are underpowered. METHODS: We queried the National Cancer Database to identify patients undergoing resection of gallbladder adenocarcinoma between 2004 and 2015. Patients presenting with metastatic disease, those who received neoadjuvant therapy, and those with fewer than 3 lymph nodes assessed were excluded. 1:1 propensity score matching was used to develop cohorts undergoing either R0 or R1 resection, matched for demographic, pathologic, and facility characteristics. Kaplan-Meier analysis was used to assess the association between margin status and overall survival. RESULTS: A total of 1,439 patients met inclusion criteria; 1,285 underwent R0 and 154 underwent R1 resection. On Kaplan-Meier analysis of propensity-matched cohorts, patients undergoing R0 resection had a median overall survival that was 18 months longer than those undergoing R1 resection (34.6 ±â€¯2.0 months vs 16.3 ±â€¯1.7 months, P < .001). CONCLUSION: In patients presenting with resectable gallbladder adenocarcinoma, margin-negative resection is associated with significant improvement in overall survival.

State-Level Examination of Clinical Outcomes and Costs for Robotic and Laparoscopic Approach to Diaphragmatic Hernia Repair.

BACKGROUND: Published studies evaluating the effect of robotic assistance on clinical outcomes and costs of care in diaphragmatic hernia repair (DHR) have been limited. STUDY DESIGN: The Healthcare Cost and Utilization Project State Inpatient and State Ambulatory Surgery and Services Databases for Florida were queried to identify patients undergoing transabdominal DHR between 2011 and 2018 and associated inpatient and outpatient encounters within 12 months after the index operation. Patients undergoing robotic DHR were 1:1:1 propensity score-matched for age, sex, race, Elixhauser comorbidity score, case priority, payer, and facility volume with patients undergoing open and laparoscopic DHR. RESULTS: There were 5,962 patients (67.3%) who underwent laparoscopic DHR, 1,520 (17.2%) who underwent open DHR, and 1,376 (15.5%) who underwent robotic DHR. On comparison of matched cohorts, median index length of stay (3 days; interquartile range [IQR] 2 to 5 days vs 2 days; IQR 1 to 4 days; p < 0.001) and index hospitalization costs ($17,236; IQR $13,231 to $22,183 vs $12,087; IQR $8,881 to $17,439; p < 0.001) for robotic DHR were greater than for laparoscopic DHR. Median length of stay for open DHR (6 days; IQR 4 to 10 days) was longer than that for both laparoscopic and robotic DHR. Median index hospitalization costs for open DHR ($16,470; IQR $11,152 to $23,768) were greater than those for laparoscopic DHR, but less than those for robotic DHR. There were no significant differences between cohorts in the overall rate of post-index care. CONCLUSIONS: Laparoscopic DHR is the most cost-effective approach to DHR. Robotic assistance provides clinical outcomes comparable with laparoscopic DHR, but is associated with increased index cost.

A Leg to Stand on: Trauma Center Designation and Association with Rate of Limb Salvage in Patients Suffering Severe Lower Extremity Injury.

BACKGROUND: Mangled extremities are one of the most difficult injuries for trauma surgeons to manage. We compare limb salvage rates for a limb-threatening lower extremity injuries managed at Level I vs Level II trauma centers (TCs). STUDY DESIGN: We identified all adult patients with a limb-threatening injury who underwent primary amputation or limb salvage (LS) using the American College of Surgeons (ACS) Trauma Quality Improvement Program database at ACS Level I vs II TCs between 2007 and 2017. A limb-threatening injury was defined as an open tibial fracture with concurrent arterial injury (Gustilo type IIIc). Multivariable analysis and propensity score matching were performed to minimize confounding by indication. RESULTS: There were 712 records for analysis; 391 (54.9%) LS performed and 321 (45.1%) underwent amputation. The rate of LS was statistically higher among patients treated at Level I TCs vs those treated at Level II TCs (47.4% vs 34.8%; p = 0.01). Patients with penetrating injuries (13% vs 9.5%; p = 0.046) and tibial/peroneal artery injury (72.9% vs 50.4%; p < 0.001), as opposed to popliteal artery injury (30.8% vs 58.8%; p < 0.001), were more likely to have LS. The risk-adjusted odds of LS was 3.13 times higher at Level I TCs vs Level II TCs (95% CI, 1.59 to 6.34; p = 0.001). Limb salvage rates were significantly higher at Level I TCs compared with Level II TCs (53.0% vs 34.8%; p = 0.004), even after propensity matching. CONCLUSIONS: In patients with a mangled extremity, limb salvage rates are 50% higher at Level I TCs compared with Level II TCs, independent of case mix and injury severity.

Does minimally invasive pancreaticoduodenectomy increase the chance of a textbook oncologic outcome?

BACKGROUND: Textbook oncologic outcome has been described in an effort to improve upon traditional outcomes defining care after pancreaticoduodenectomy for adenocarcinoma. We sought to examine whether minimally invasive pancreaticoduodenectomy increased the likelihood of an optimal textbook oncologic outcome. METHODS: Patients undergoing open pancreaticoduodenectomy or minimally invasive pancreaticoduodenectomy between 2010 and 2015 were identified in the National Cancer Database. Textbook oncologic outcome was defined as R0 resection with American Joint Committee on Cancer compliant lymphadenectomy, no prolonged duration of stay, no 30-day readmission/mortality, and receipt of adjuvant chemotherapy. Propensity score matching was employed to evaluate adjusted rates of textbook oncologic outcome, in addition to overall survival. RESULTS: Among 12,854 patients who underwent pancreaticoduodenectomy, 48.3% were female, and the median patient age was 66 years; 87.5% underwent open pancreaticoduodenectomy, and 12.5% underwent whether minimally invasive pancreaticoduodenectomy. After propensity score matching, there were no noted differences in the likelihood of R0 resection, adequate lymphadenectomy, nonprolonged duration of stay, no readmission, no 30-day mortality, adjuvant chemotherapy, or textbook oncologic outcome among open pancreaticoduodenectomy versus minimally invasive pancreaticoduodenectomy (P > .05). Textbook oncologic outcome was associated with an improved median overall survival (negative textbook oncologic outcome: 21.3 months vs positive textbook oncologic outcome: 27.6 months, P < .001). CONCLUSION: Although textbook oncologic outcome was associated with a survival advantage, it was only achieved in 1 in 4 patients undergoing pancreaticoduodenectomy for adenocarcinoma. Achievement of textbook oncologic outcome was equivalent among patients who underwent minimally invasive pancreaticoduodenectomy compared with open pancreaticoduodenectomy after propensity score matching.

Comparison and interpretability of machine learning models to predict severity of chest injury.

OBJECTIVE: Trauma quality improvement programs and registries improve care and outcomes for injured patients. Designated trauma centers calculate injury scores using dedicated trauma registrars; however, many injuries arrive at nontrauma centers, leaving a substantial amount of data uncaptured. We propose automated methods to identify severe chest injury using machine learning (ML) and natural language processing (NLP) methods from the electronic health record (EHR) for quality reporting. MATERIALS AND METHODS: A level I trauma center was queried for patients presenting after injury between 2014 and 2018. Prediction modeling was performed to classify severe chest injury using a reference dataset labeled by certified registrars. Clinical documents from trauma encounters were processed into concept unique identifiers for inputs to ML models: logistic regression with elastic net (EN) regularization, extreme gradient boosted (XGB) machines, and convolutional neural networks (CNN). The optimal model was identified by examining predictive and face validity metrics using global explanations. RESULTS: Of 8952 encounters, 542 (6.1%) had a severe chest injury. CNN and EN had the highest discrimination, with an area under the receiver operating characteristic curve of 0.93 and calibration slopes between 0.88 and 0.97. CNN had better performance across risk thresholds with fewer discordant cases. Examination of global explanations demonstrated the CNN model had better face validity, with top features including "contusion of lung" and "hemopneumothorax." DISCUSSION: The CNN model featured optimal discrimination, calibration, and clinically relevant features selected. CONCLUSION: NLP and ML methods to populate trauma registries for quality analyses are feasible.

The Addition of United States Census-Tract Data Does Not Improve the Prediction of Substance Misuse.

Predictors from the structured data in the electronic health record (EHR) have previously been used for case-identification in substance misuse. We aim to examine the added benefit from census-tract data, a proxy for socioeconomic status, to improve identification. A cohort of 186,611 hospitalizations was derived between 2007 and 2017. Reference labels included alcohol misuse only, opioid misuse only, and both alcohol and opioid misuse. Baseline models were created using 24 EHR variables, and enhanced models were created with the addition of 48 census-tract variables from the United States American Community Survey. The absolute net reclassification index (NRI) was applied to measure the benefit in adding census-tract variables to baseline models. The baseline models already had good calibration and discrimination. Adding census-tract variables provided negligible improvement to sensitivity and specificity and NRI was less than 1% across substance groups. Our results show the census-tract added minimal value to prediction models.