Postoperative Outcomes Among Patients Undergoing Cancer Surgery: United States versus International Medical Graduates.

OBJECTIVE: We sought to characterize postoperative outcomes among patients who underwent an oncologic operation relative to whether the treating surgeon was an international medical graduate (IMG) versus a United States medical graduate (USMG). BACKGROUND: IMGs comprise approximately one quarter of the physician workforce in the United States. METHODS: The 100% Medicare Standard Analytic Files were utilized to extract data on patients with breast, lung, hepato-pancreato-biliary (HPB), and colorectal cancer who underwent surgical resection between 2014 and 2020. Entropy balancing and multivariable regression analysis were performed to evaluate the association between postoperative outcomes among USMG and IMG surgeons. RESULTS: Among 285,930 beneficiaries, 242,914 (85.0%) and 43,016 (15.0%) underwent surgery by a USMG or IMG surgeon, respectively. Overall, 129,576 (45.3%) individuals were male, and 168,848 (59.1%) patients had a Charlson Comorbidity Index score >2. Notably, IMG surgeons were more likely to care for racial/ethnic minority patients (14.7% vs 12.5%) and individuals with a high social vulnerability index (33.3% vs 32.1%) (all P <0.001). On multivariable analysis after entropy balancing, patients treated by an IMG surgeon were less likely to experience adverse postoperative outcomes, including 90-day readmission [odds ratio (OR) 0.89, 95% CI: 0.80-0.99] and index complications (OR: 0.84, 95% CI: 0.74-0.95) versus USMG surgeons (all P <0.05). Patients treated by IMG versus USMG surgeons had no difference in likelihood to achieve a textbook outcome (OR: 1.10, 95% CI: 0.99-1.21; P =0.077). CONCLUSIONS: Postoperative outcomes among patients treated by IMG surgeons were roughly equivalent to those of USMG surgeons. In addition, IMG surgeons were more likely to care for patients with multiple comorbidities and individuals from vulnerable communities.

Practice Patterns and Outcomes Among Surgical Oncology Fellowship Graduates Performing Complex Cancer Surgery in the United States Across Different Career Stages.

BACKGROUND: Practice patterns and potential quality differences among surgical oncology fellowship graduates relative to years of independent practice have not been defined. METHODS: Medicare claims were used to identify patients who underwent esophagectomy, pancreatectomy, hepatectomy, or rectal resection for cancer between 2016 and 2021. Surgical oncology fellowship graduates were identified, and the association between years of independent practice, serious complications, and 90-day mortality was examined. RESULTS: Overall, 11,746 cancer operations (pancreatectomy [61.2%], hepatectomy [19.5%], rectal resection [13.7%], esophagectomy [5.6%]) were performed by 676 surgical oncology fellowship graduates (females: 17.7%). The operations were performed for 4147 patients (35.3%) by early-career surgeons (1-7 years), for 4104 patients (34.9%) by mid-career surgeons (8-14 years), and for 3495 patients (29.8%) by late-career surgeons (>15 years). The patients who had surgery by early-career surgeons were treated more frequently at a Midwestern (24.9% vs. 14.2%) than at a Northeastern institution (20.6% vs. 26.9%) compared with individuals treated by late-career surgeons (p < 0.05). Surgical oncologists had comparable risk-adjusted serious complications and 90-day mortality rates irrespective of career stage (early career [13.0% and 7.2%], mid-career [12.6% and 6.3%], late career [12.8% and 6.5%], respectively; all p > 0.05). Surgeon case-specific volume independently predicted serious complications across all career stages (high vs. low volume: early career [odds ratio {OR}, 0.80; 95% confidence interval {CI}, 0.65-0.98]; mid-career [OR, 0.81; 95% CI, 0.66-0.99]; late career [OR, 0.78; 95% CI, 0.62-0.97]). CONCLUSION: Among surgical oncology fellowship graduates performing complex cancer surgery, rates of serious complications and 90-day mortality were comparable between the early-career and mid/late-career stages. Individual surgeon case-specific volume was strongly associated with postoperative outcomes irrespective of years of independent practice or career stage.

Clinical Outcomes, Costs, and Value of Surgery Among Older Patients with Colon Cancer at US News and World Report Ranked Versus Unranked Hospitals.

BACKGROUND: US News and World Report (USNWR) hospital rankings influence patient choice of hospital, but their association with surgical outcomes remains ill-defined. We sought to characterize clinical outcomes and costs of surgery for colon cancer among USNWR top ranked and unranked hospitals. METHODS: Using Medicare Standard Analytic Files, patients aged ≥65 years undergoing surgery for colon cancer were identified. Hospitals were categorized as 'ranked' or 'unranked' based on USNWR cancer hospital rankings. One-to-one matching was performed between patients treated at ranked and unranked hospitals, and clinical outcomes and costs of surgery were compared. RESULTS: Among 50 ranked and 2522 unranked hospitals, 13,650 patient pairs were compared. Overall, 30-day mortality was 2.13% in ranked hospitals versus 3.68% in unranked hospitals (p < 0.0001), and the overall paired cost difference was $8159 (p < 0.0001). As patient risk increased, 30-day mortality differences became larger, with the ranked hospitals having 30-day mortality of 7.59% versus 11.84% for unranked hospitals among the highest-risk patients (p < 0.0001). Overall paired cost differences also increased with increasing patient risk, with cost of care being $72,229 for ranked hospitals versus $56,512 for unranked hospitals among the highest-risk patients (difference = $14,394; p = 0.02). The difference in cost per 1% reduction in 30-day mortality was $9009 (95% confidence interval [CI] $6422-$11,597) for lowest-risk patients, which dropped to $3387 (95% CI $2656-$4119) for highest-risk patients (p < 0.0001). CONCLUSION: Treatment at USNWR-ranked hospitals, particularly for higher-risk patients, was associated with better outcomes but higher-cost care. The benefit of being treated at highly ranked USNWR hospitals was most pronounced among high-risk patients.

Long-term Health Outcomes of New Persistent Opioid Use After Gastrointestinal Cancer Surgery.

BACKGROUND: New persistent opioid use (NPOU) after surgery has been identified as a common complication. This study sought to assess the long-term health outcomes among patients who experienced NPOU after gastrointestinal (GI) cancer surgery. METHODS: Patients who underwent surgery for hepato-pancreato-biliary and colorectal cancer between 2007 and 2019 were identified using the Surveillance, Epidemiology, and End Results (SEER)-Medicare-linked database. Mixed-effect multivariable logistic regression and Cox proportional hazard models were used to estimate the risk of mortality and hospital visits related to falls, respiratory events, or pain symptoms. RESULTS: Among 15,456 patients who underwent GI cancer surgery, 967(6.6%) experienced NPOU. Notably, the patients at risk for the development of NPOU were those with a history of substance abuse (odds ratio [OR], 1.45; 95% confidence interval [CI], 1.14-1.84), moderate social vulnerability (OR, 1.26; 95% CI, 1.06-1.50), an advanced disease stage (OR, 4.42; 95% CI, 3.51-5.82), or perioperative opioid use (OR, 3.07; 95% CI, 2.59-3.63. After control for competing risk factors, patients who experienced NPOU were more likely to visit a hospital for falls, respiratory events, or pain symptoms (OR, 1.45, 95% CI 1.18-1.78). Moreover, patients who experienced NPOU had a greater risk of death at 1 year (hazard ratio [HR], 2.15; 95% CI, 1.74-2.66). CONCLUSION: Approximately 1 in 15 patients experienced NPOU after GI cancer surgery. NPOU was associated with an increased risk of subsequent hospital visits and higher mortality. Targeted interventions for individuals at higher risk for NPOU after surgery should be used to help mitigate the harmful effects of NPOU.

Impact of GLP-1RA on the Risk of Adverse Liver Outcomes Among Patients With Alcohol-Associated Liver Disease and Type 2 Diabetes.

BACKGROUND AND AIMS: We sought to characterise the impact of GLP-1RA on adverse liver outcomes (ALO) among patients with alcohol-associated liver disease (ALD) and Type 2 diabetes mellitus (T2DM). METHODS: Patients with T2DM newly diagnosed with ALD between 2013 and 2020 were identified using IBM MarketScan database and were categorised by GLP-1RA exposure. Overlap propensity score weighting (OPSW) followed by Poisson regression models was used to analyse adjusted risk of ALO, a composite endpoint defined by first occurrence of hepatic decompensation (HD), portal hypertension (PH), hepatocellular carcinoma (HCC) or liver transplantation (LT) relative to GLP-1RA. RESULTS: Among 14 730 patients, most individuals were male (n = 9752, 66.2%) with median age of 57 (IQR 52-61) years; 2.2% (n = 317) of patients had GLP-1RA exposure. Overall, 32.0% (n = 4717) of patients experienced HD, 15.9% (n = 2345) had PH, 3.8% (n = 563) developed HCC, while 2.5% (n = 374) underwent transplantation. Non-GLP-1RA patients had higher incidence of HD (32.2% vs. 22.4%) and HCC (3.9% vs. 0.3%) versus patients taking GLP-1RA (both p < 0.001); in contrast, there was no difference in incidence of PH (14.5% vs. 16.0%) and LT (1.3% vs. 2.6%) (both p > 0.05). After OPSW, overall incidence of ALO was lower in GLP-1RA cohort (GLP-1RA: 12.0%, 95%CI 9.0-16.0 vs. non-GLP-1RA: 21.0%, 95%CI 20.0-22.0) with an absolute incidence risk reduction of 9.0% (95%CI 3.0%-15.0%) associated with GLP-1RA. GLP-1RA was most strongly associated with lower likelihood of HD with reduced adjusted incidence rate of 0.56 (95%CI 0.36-0.86) relative to non-GLP-1RA individuals. CONCLUSIONS: GLP-1RA may have a hepatoprotective impact among patients with ALD and T2DM.

The Association of Socio-Environmental Inequality and Outcomes Among Patients Undergoing Major Surgery.

INTRODUCTION: Environmental hazards may influence health outcomes and be a driver of health inequalities. We sought to characterize the extent to which social-environmental inequalities were associated with surgical outcomes following a complex operation. METHODS: In this cross-sectional study, patients who underwent abdominal aortic aneurysm repair, coronary artery bypass grafting, colectomy, pneumonectomy, or pancreatectomy between 2016 and 2021 were identified from Medicare claims data. Patient data were linked with social-environmental data sourced from Centers for Disease Control and Agency for Toxic Substances and Disease Registry data based on county of residence. The Environmental Justice Index social-environmental ranking (SER) was used as a measure of environmental injustice. Multivariable regression analysis was performed to assess the relationship between SER and surgical outcomes. RESULTS: Among 1,052,040 Medicare beneficiaries, 346,410 (32.9%) individuals lived in counties with low SER, while 357,564 (33.9%) lived in counties with high SER. Patients experiencing greater social-environmental injustice were less likely to achieve textbook outcome (odds ratio 0.95, 95% confidence interval 0.94-0.96, P < 0.001) and to be discharged to an intermediate care facility or home with a health agency (odds ratio 0.97, 95% confidence interval 0.96-0.98, P < 0.001). CONCLUSIONS: Cumulative social and environmental inequalities, as captured by the Environmental Justice Index SER, were associated with postoperative outcomes among Medicare beneficiaries undergoing a range of surgical procedures. Policy makers should focus on environmental, as well as socioeconomic injustice to address preventable health disparities.

Association of surgeon's sex, and surgeon-patient dyad with financial outcomes among patients undergoing cancer surgery.

BACKGROUND AND OBJECTIVES: Surgeon sex has been associated with perioperative clinical outcomes among patients undergoing oncologic surgery. There may be variations in financial outcomes relative to the surgeon-patient dyad. We sought to define the association of surgeon's sex with perioperative financial outcomes following cancer surgery. METHODS: Patients who underwent resection of lung, breast, hepato-pancreato-biliary (HPB), or colorectal cancer between 2014 and 2021 were identified from the Medicare Standard Analytic Files. A generalized linear model with gamma regression was utilized to characterize the association between sex concordance and expenditures. RESULTS: Among 207,935 Medicare beneficiaries (breast: n = 14,753, 7.1%, lung: n = 59,644, 28.7%, HPB: n = 23,400, 11.3%, colorectal: n = 110,118, 53.0%), 87.8% (n = 182,643) and 12.2% (n = 25,292) of patients were treated by male and female surgeons, respectively. On multivariable analysis, female surgeon sex was associated with slightly reduced index expenditures (mean difference -$353, 95%CI -$580, -$126; p = 0.003). However, there were no differences in 90-day post-discharge inpatient (mean difference -$-225, 95%CI -$570, -$121; p = 0.205) and total expenditures (mean difference $133, 95%CI -$279, $545; p = 0.525). CONCLUSIONS: There was minor risk-adjusted variation in perioperative expenditures relative to surgeon sex. To improve perioperative financial outcomes, a diverse surgical workforce with respect to patient and surgeon sex is warranted.

Impact of Spousal Mental Illness on Healthcare Utilization Among Patients With Gastrointestinal Cancer.

BACKGROUND: Informal caregiving involves increased responsibilities, with financial and emotional challenges, thereby affecting the well-being of the caregiver. We aimed to investigate the effect of spousal mental illness on hospital visits and medical spending among patients with gastrointestinal (GI) cancer. METHODS: Patients who underwent GI cancer surgery between 2013 and 2020 were identified from the IBM Marketscan database. Multivariable regression analysis was used to examine the association between spousal mental illness and healthcare utilization. RESULTS: A total of 6,035 patients underwent GI surgery for a malignant indication. Median age was 54 years (IQR: 49-59), most patients were male (n = 3592, 59.5%), and had a CCI score of ≤ 2 (n = 5512, 91.3%). Of note, in the 1 year follow-up period, 19.4% (anxiety: n = 509, 8.4%; depression: n = 301, 5.0%; both anxiety and depression: n = 273, 4.5%; severe mental illness: n = 86, 1.4%) of spouses developed a mental illness. On multivariable analysis, after controlling for competing factors, spousal mental illness remained independently associated with increased odds of emergency department visits (OR 1.20, 95% CI 1.05-1.38) and becoming a super healthcare utilizer (OR 1.37, 95% CI 1.04-1.79), as well as 12.1% (95% CI 10.6-15.3) higher medical spending. CONCLUSION: Among patients with GI cancer spousal mental illness is associated with higher rates of outpatient visits, emergency department visits, and expenditures during the 1-year postoperative period. These findings underscore the importance of caregiving resources and counseling in alleviating caregiver burden, thereby reducing the overall burden on the healthcare system.

Impact of early primary care physician follow-up on hospital readmission following gastrointestinal cancer surgery.

BACKGROUND: We sought to examine the association between primary care physician (PCP) follow-up on readmission following gastrointestinal (GI) cancer surgery. METHODS: Patients who underwent surgery for GI cancer were identified using the Surveillance, Epidemiology and End Results (SEER) database. Multivariable regression was performed to examine the association between early PCP follow-up and hospital readmission. RESULTS: Among 60 957 patients who underwent GI cancer surgery, 19 661 (32.7%) visited a PCP within 30-days after discharge. Of note, patients who visited PCP were less likely to be readmitted within 90 days (PCP visit: 17.4% vs. no PCP visit: 28.2%; p < 0.001). Median postsurgical expenditures were lower among patients who visited a PCP (PCP visit: $4116 [IQR: $670-$13 860] vs. no PCP visit: $6700 [IQR: $870-$21 301]; p < 0.001). On multivariable analysis, PCP follow-up was associated with lower odds of 90-day readmission (OR: 0.52, 95% CI: 0.50-0.55) (both p < 0.001). Moreover, patients who followed up with a PCP had lower risk of death at 90-days (HR: 0.50, 95% CI: 0.40-0.51; p < 0.001). CONCLUSION: PCP follow-up was associated with a reduced risk of readmission and mortality following GI cancer surgery. Care coordination across in-hospital and community-based health platforms is critical to achieve optimal outcomes for patients.

Surgical outcomes and healthcare expenditures among patients with dementia undergoing major surgery.

BACKGROUND: We sought to define surgical outcomes among elderly patients with Alzheimer's disease and related dementias (ADRD) following major thoracic and gastrointestinal surgery. METHODS: A retrospective cohort study was used to identify patients who underwent coronary artery bypass grafting (CABG), abdominal aortic aneurysm (AAA) repair, pneumonectomy, pancreatectomy, and colectomy. Individuals were identified from the Medicare Standard Analytic Files and multivariable regression was utilized to assess the association of ADRD with textbook outcome (TO), expenditures, and discharge disposition. RESULTS: Among 1,175,010 Medicare beneficiaries, 19,406 (1.7%) patients had a preoperative diagnosis of ADRD (CABG: n = 1,643, 8.5%; AAA repair: n = 5,926, 30.5%; pneumonectomy: n = 590, 3.0%; pancreatectomy: n = 181, 0.9%; and colectomy: n = 11,066, 57.0%). After propensity score matching, patients with ADRD were less likely to achieve a TO (ADRD: 31.2% vs. no ADRD: 40.1%) or be discharged to home (ADRD: 26.7% vs. no ADRD: 46.2%) versus patients who did not have ADRD (both p < 0.001). Median index surgery expenditures were higher among patients with ADRD (ADRD: $28,815 [IQR $14,333-$39,273] vs. no ADRD: $27,101 [IQR $13,433-$38,578]; p < 0.001) (p < 0.001). On multivariable analysis, patients with ADRD had higher odds of postoperative complications (OR 1.32, 95% CI 1.25-1.40), extended length-of-stay (OR 1.26, 95% CI 1.21-1.32), 90-day readmission (OR 1.37, 95% CI 1.31-1.43), and 90-day mortality (OR 1.76, 95% CI 1.66-1.86) (all p < 0.001). CONCLUSION: Preoperative diagnosis of ADRD was an independent risk factor for poor postoperative outcomes, discharge to non-home settings, as well as higher healthcare expenditures. These data should serve to inform discussions and decision-making about surgery among the growing number of older patients with cognitive deficits.

Association of county-level provider density with hepatobiliary cancer incidence and mortality.

BACKGROUND: Access to healthcare providers is a key factor in reducing cancer incidence and mortality, underscoring the significance of provider density as a crucial metric of health quality. We sought to characterize the association of provider density on hepatobiliary cancer population-level incidence and mortality. STUDY DESIGN: County-level hepatobiliary cancer incidence and mortality data from 2016 to 2020 and provider data from 2016 to 2018 were obtained from the CDC and Area Health Resource File. Multivariable logistic regression was utilized to evaluate the relationship between provider density and hepatobiliary cancer incidence and mortality. RESULTS: Among 1359 counties, 851 (62.6%) and 508 (37.4%) counties were categorized as urban and rural, respectively. The median number of providers in any given county was 104 (IQR: 44-306), while provider density was 120.1 (IQR: 86.7-172.2) per 100,000 population; median household income was $51,928 (IQR: $45,050-$61,655). Low provider-density counties were more likely to have a greater proportion of residents over 65 years of age (52.7% vs. 49.6%) who were uninsured (17.4% vs. 13.2%) versus higher provider-density counties (p < 0.05). Moreover, all-stage incidence, late-stage incidence, and mortality rates were higher in counties with low provider density. On multivariable analysis, moderate, and high provider density were associated with lower odds of all-stage incidence, late-stage incidence, and mortality. CONCLUSION: Higher county-level provider density was associated with lower hepatobiliary cancer-related incidence and mortality. Efforts to increase access to healthcare providers may improve healthcare equity as well as long-term cancer outcomes.

Preoperative identification of early extrahepatic recurrence after hepatectomy for colorectal liver metastases: A machine learning approach.

BACKGROUND: Machine learning (ML) may provide novel insights into data patterns and improve model prediction accuracy. The current study sought to develop and validate an ML model to predict early extra-hepatic recurrence (EEHR) among patients undergoing resection of colorectal liver metastasis (CRLM). METHODS: Patients with CRLM who underwent curative-intent resection between 2000 and 2020 were identified from an international multi-institutional database. An eXtreme gradient boosting (XGBoost) model was developed to estimate the risk of EEHR, defined as extrahepatic recurrence within 12 months after hepatectomy, using clinicopathological factors. The relative importance of factors was determined using Shapley additive explanations (SHAP) values. RESULTS: Among 1410 patients undergoing curative-intent resection, 131 (9.3%) patients experienced EEHR. Median OS among patients with and without EEHR was 35.4 months (interquartile range [IQR] 29.9-46.7) versus 120.5 months (IQR 97.2-134.0), respectively (p < 0.001). The ML predictive model had c-index values of 0.77 (95% CI, 0.72-0.81) and 0.77 (95% CI, 0.73-0.80) in the entire dataset and the validation data set with bootstrapping resamples, respectively. The SHAP algorithm demonstrated that T and N primary tumor categories, as well as tumor burden score were the three most important predictors of EEHR. An easy-to-use risk calculator for EEHR was developed and made available online at: https://junkawashima.shinyapps.io/EEHR/. CONCLUSIONS: An easy-to-use online calculator was developed using ML to help clinicians predict the chance of EEHR after curative-intent resection for CRLM. This tool may help clinicians in decision-making related to treatment strategies for patients with CRLM.

Impact of an artificial intelligence based model to predict non-transplantable recurrence among patients with hepatocellular carcinoma.

OBJECTIVE: We sought to develop Artificial Intelligence (AI) based models to predict non-transplantable recurrence (NTR) of hepatocellular carcinoma (HCC) following hepatic resection (HR). METHODS: HCC patients who underwent HR between 2000-2020 were identified from a multi-institutional database. NTR was defined as recurrence beyond Milan Criteria. Different machine learning (ML) and deep learning (DL) techniques were used to develop and validate two prediction models for NTR, one using only preoperative factors and a second using both preoperative and postoperative factors. RESULTS: Overall, 1763 HCC patients were included. Among 877 patients with recurrence, 364 (41.5%) patients developed NTR. An ensemble AI model demonstrated the highest area under ROC curves (AUC) of 0.751 (95% CI: 0.719-0.782) and 0.717 (95% CI:0.653-0.782) in the training and testing cohorts, respectively which improved to 0.858 (95% CI: 0.835-0.884) and 0.764 (95% CI: 0.704-0.826), respectively after incorporation of postoperative pathologic factors. Radiologic tumor burden score and pathological microvascular invasion were the most important preoperative and postoperative factors, respectively to predict NTR. Patients predicted to develop NTR had overall 1- and 5-year survival of 75.6% and 28.2%, versus 93.4% and 55.9%, respectively, among patients predicted to not develop NTR (p < 0.0001). CONCLUSION: The AI preoperative model may help inform decision of HR versus LT for HCC, while the combined AI model can frame individualized postoperative care (https://altaf-pawlik-hcc-ntr-calculator.streamlit.app/).

Artificial intelligence-based model for the recurrence of hepatocellular carcinoma after liver transplantation.

BACKGROUND: Artificial intelligence-based models might improve patient selection for liver transplantation in hepatocellular carcinoma. The objective of the current study was to develop artificial intelligence-based deep learning models and determine the risk of recurrence after living donor liver transplantation for hepatocellular carcinoma. METHODS: The study was a single-center retrospective cohort study. Patients who underwent living donor liver transplantation for hepatocellular carcinoma were divided into training and validation cohorts (n = 192). The deep learning models were used to stratify patients in the training cohort into low- and high-risk groups, and 5-year recurrence-free survival was assessed in the validation cohort. RESULTS: The median follow-up period was 59.1 (33.9-72.4) months. The artificial intelligence model (pretransplant factors) had an area under the curve of 0.86 in the training cohort and 0.71 in the validation cohort. The largest tumor diameter and alpha-fetoprotein level had the greatest Shapley Additive exPlanations values for recurrence (>0.4). The 5-year recurrence-free survival rates in the low- and high-risk groups were 92.6% and 45% (P < .001). In the second artificial intelligence model (pretransplant factors + grade), the area under the curve for the validation cohort was 0.77, with 5-year recurrence-free survival rates of 96% and 30% in the low- and high-risk groups (P < .001). None of the low-risk patients outside the Milan and University of California San Francisco Criteria had recurrence during follow-up. CONCLUSIONS: The artificial intelligence-based hepatocellular carcinoma transplant recurrence models might improve patient selection for liver transplantation.

Upper gastrointestinal cancers: Trends and determinants of location of death.

BACKGROUND: Patients diagnosed with upper gastrointestinal cancers often require extensive end-of-life care. We sought to investigate social determinants of health associated with disparities in the location of death among patients who died of upper gastrointestinal cancers. METHODS: Patients who died between 2003 and 2020 from esophageal cancer, gastric cancer, hepatobiliary cancer, and pancreatic cancer were identified using the Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research database. Social determinants of health were assessed using the Social Vulnerability Index. Patients were categorized on the basis of location of death: inpatient hospital, home, nursing home, hospice, and outpatient medical facility/emergency department. Multivariable regression and mediation analyses defined the association of patient race as well as social determinants of health with location of death. RESULTS: Among 815,780 decedents (esophageal cancer: 15.3%; gastric cancer: 3.6%; hepatobiliary cancer: 36.6%; pancreatic cancer: 54.5%), most were male (60.8%), aged 55-74 years (52.3%), and White (89.1%). Most decedents died at home (55.7%), followed by inpatient hospital (24.8%), hospice (9.0%), nursing home (8.1%), and outpatient medical facility/emergency department (2.5%). During the study period, location of death shifted notably from inpatient hospital (36.8% to 21.3%) to home (45.8% to 56.3%). Residents of high Social Vulnerability Index areas were more likely to die at inpatient hospital compared with home (31.8% vs 24.3%) (P < .001). Black race (reference: White; odds ratio; 0.41, 95% confidence interval, 0.40-0.42) and social vulnerability (reference: low Social Vulnerability Index; odds ratio, 0.64, 95% confidence interval, 0.63-0.65) remained independently associated with lower odds of dying at home compared with an inpatient hospital. Notably, 65% of the overall race-based association with death at inpatient hospital was driven indirectly through social determinants of health. CONCLUSION: Social determinants are important drivers of end-of-life care and impact the potential ability of patients with cancer to die at home.

Disparities in clinical trial enrollment among patients with gastrointestinal cancer relative to minority-serving and safety-netting hospitals.

BACKGROUND: For results to be generalizable to all patients with cancer, clinical trials need to include a diverse patient demographic that is representative of the general population. We sought to characterize the effect of receiving care at a minority-serving hospital (MSH) and/or safety-net hospital on clinical trial enrollment among patients with gastrointestinal (GI) malignancies. METHODS: Adult patients with GI cancer who underwent oncologic surgery and were enrolled in institutional-/National Cancer Institute-funded clinical trials between 2012 and 2019 were identified in the National Cancer Database. Multivariable regression was used to assess the relationship between MSH and safety-net status relative to clinical trial enrollment. RESULTS: Among 1,112,594 patients, 994,598 (89.4%) were treated at a non-MSH, whereas 117,996 (10.6%) were treated at an MSH. Only 1857 patients (0.2%) were enrolled in a clinical trial; most patients received care at a non-MSH (1794 [96.6%]). On multivariable analysis, the odds of enrollment in a clinical trial were markedly lower among patients treated at an MSH vs non-MSH (odds ratio [OR], 0.32; 95% CI, 0.22-0.46). In addition, even after controlling for receipt of care at MSH, Black patients remained at lower odds of enrollment in a clinical trial than White patients (OR, 0.57; 95% CI, 0.45-0.73; both P < .05). CONCLUSION: Overall, clinical trial participation among patients with GI cancer was extremely low. Patients treated at an MSH and high safety-net burden hospitals and Black individuals were much less likely to be enrolled in a clinical trial. Efforts should be made to improve trial enrollment and address disparities in trial representation.

Impact of patient, hospital, and operative characteristics relative to social determinants of health: Compliance with National Comprehensive Cancer Network guidelines for colon cancer.

BACKGROUND: Despite an established association with improved patient outcomes, compliance with National Comprehensive Cancer Network (NCCN) guidelines remains suboptimal. We sought to assess the effect of patient characteristics (PCs), operative characteristics (OCs), hospital characteristics (HCs), and social determinants of health (SDoH) on noncompliance with NCCN guidelines for colon cancer. METHODS: Patients treated for stage I to III colon cancer from 2004 to 2017 were identified from the National Cancer Database. Multilevel multivariate regression analysis was performed to identify factors associated with receipt of NCCN-compliant care and quantify the proportion of variance explained by PCs, OCs, HCs, and SDoH. RESULTS: Among 468,097 patients with colon cancer treated across 1319 hospitals, 1 in 4 patients did not receive NCCN-compliant care (122,170 [26.1%]). On regression analysis, older age (odds ratio [OR], 0.96; 95% CI, 0.96-0.96), female sex (OR, 0.97; 95% CI, 0.96-0.99), Black race (OR, 0.96; 95% CI, 0.94-0.98), higher Charlson-Deyo score (OR, 0.84; 95% CI, 0.82-0.86), tumor stage ≥II (OR, 0.42; 95% CI, 0.40-0.44), and tumor grade ≥ 3 (OR, 0.33; 95% CI, 0.32-0.34) were associated with lower odds of receiving NCCN-compliant care (all P values <.05). Higher hospital volume (OR, 1.02; 95% CI, 1.02-1.03), minimally invasive or robotic surgical approach (OR, 1.26; 95% CI, 1.23-1.29), adequate (≥12) lymph node assessment (OR, 3.46; 95% CI, 3.38-3.53), private insurance status (OR, 1.33; 95% CI, 1.26-1.40), Medicare insurance status (OR, 1.42; 95% CI, 1.35-1.49), and higher educational status (OR, 1.06; 95% CI, 1.02-1.09) were associated with higher odds of receiving NCCN-compliant care (all P values <.05). Overall, PCs contributed 36.5%, HCs contributed 1.3%, and OCs contributed 12.9% to the variation in guideline-compliant care, while SDoH contributed only 3.6% of the variation in receipt of NCCN-compliant care. CONCLUSION: The variation in NCCN-compliant care among patients with colon cancer was largely attributable to patient- and surgeon-level factors, whereas SDoH were associated with a smaller proportion of the variation.

Variation in physician spending and its association with postoperative outcomes among patients undergoing surgery for gastrointestinal cancer.

BACKGROUND: There is significant variation in inpatient expenditures among physicians and hospitals. This study aimed to characterize the association between variation in physician spending and short-term outcomes among patients undergoing surgery for pancreatic ductal adenocarcinoma (PDAC) and colorectal cancer (CRC). METHODS: Patients who underwent surgery for PDAC and CRC from 2010 to 2020 were identified using the Surveillance, Epidemiology, and End Result-Medicare-linked database. The cohort was divided into quartiles based on adjusted physician spending, and multivariate models were used to assess the association between physician spending and patient outcomes. RESULTS: Among 27,596 Medicare beneficiaries, 25,615 (92.8%) underwent surgery for CRC and 1981 (7.2%) underwent surgery for PDAC. Of the variations in spending, 79.9% were due to patient-level factors, 13.3% were due to hospital characteristics, and 6.8% were due to surgeon-level variables. On multivariate analysis, there was no association between physician spending and 30-day readmission (with complications: first quartile [Q1], reference; Q4: odds ratio [OR], 1.10; 95% CI, 0.86-1.41; P = .123; without complications: Q1, reference; Q3, stage IV: OR, 0.97; 95% CI, 0.68-1.40; P = .882) or between physician spending and 30-day mortality (without complications: Q1, reference; Q2, stage I: OR, 1.17; 95% CI, 0.45-3.01; P = .804). However, an increase in physician spending was associated with higher 30-day mortality among patients with complications (Q1, reference; Q4: OR, 2.28; 95% CI, 1.72-3.03; P < .001). CONCLUSION: There was more variation in healthcare spending across hospitals than across individual physicians. No consistent association between variation in physician spending and patient outcomes was noted. Wasteful spending can be reduced through targeted interventions aimed at reducing variations at the physician and hospital levels.

Healthcare utilization and expenditures among patients with venous thromboembolism following gastrointestinal cancer surgery.

BACKGROUND: We sought to assess healthcare utilization and expenditures among patients who developed venous thromboembolism (VTE) after gastrointestinal cancer surgery. METHODS: Patients who underwent surgery for esophageal, gastric, hepatic, biliary duct, pancreatic, and colorectal cancer between 2013 and 2020 were identified using the MarketScan database. Entropy balancing was performed to obtain a cohort that was well balanced relative to different clinical covariates. Generalized linear models were used to compare 1-year postdischarge costs among patients who did and did not develop a postoperative VTE. RESULTS: Among 20,253 individuals in the analytical cohort (esophagus [n = 518 {2.6%}], stomach [n = 970 {4.8%}], liver [n = 608 {3.0%}], bile duct [n = 294 {1.5%}], pancreas [n = 1511 {7.5%}], colon [n = 12,222 {60.3%}], and rectum [n = 4130 {20.4%}]), 894 (4.4%) developed VTE. Overall, most patients were male (n = 10,656 [52.6%]), aged between 55 and 64 years (n = 10,372 [51.2%]), and were employed full time (n = 11,408 [56.3%]). On multivariable analysis, VTE was associated with higher inpatient (mean difference [MD], $17,547; 95% CI, $15,141-$19,952), outpatient (MD, $8769; 95% CI, $7045-$10,491), and pharmacy (MD, $2811; 95% CI, $2509-$3113) expenditures (all P < .001). Furthermore, patients who developed VTE had higher out-of-pocket costs for inpatient (MD, $159; 95% CI, $66-$253) and pharmacy (MD, $122; 95% CI, $109-$136) services (all P < .001). CONCLUSION: Among privately insured patients aged <65 years, VTE was associated with increased healthcare utilization and expenditures during the first year after discharge.