Variation in the cost-of-rescue among medicare patients with complications following hepatopancreatic surgery

Care fragmentation in hepatopancreatic surgery and postoperative outcomes

BACKGROUND

Practice fragmentation in surgery may be associated with poor quality of care. We sought to define the association between fragmented practice and outcomes in hepatopancreatic surgery relative to surgeon volume and sex.

METHODS

Medicare beneficiaries who underwent hepatopancreatic surgery between 2016 and 2021 were identified. Multivariable analysis was performed to determine provider sex-based differences in the rate of fragmented practice relative to the achievement of a textbook outcome and health care expenditures after adjusting for procedure-specific case volume.

RESULTS

Among 37,416 patients, almost one-half were female (n = 18,333, 49.0%) with the majority treated by male surgeons (n = 33,697, 90.8%). Female surgeons were more likely to have a greater rate of fragmented practice (females: n = 242, 84.9% vs males: n = 1,487, 78.4%, P = .003; odds ratio 2.66, 95% confidence interval 2.33-3.03, P < .001). Patients treated by high rate of fragmented practice surgeons had increased odds of postoperative complications (odds ratio 1.40, 95% confidence interval 1.28-1.54), extended length-of-stay (odds ratio 1.52, 95% confidence interval 1.38-1.68), 90-day-mortality (odds ratio 1.49, 95% confidence interval 1.28-1.72), and lower odds of achieving a textbook outcome (odds ratio 0.76, 95% confidence interval 0.71-0.83). This association persisted independent of surgeon-specific volume (textbook outcome, high vs low rate of fragmented practice: high-volume surgeon, odds ratio 0.53, 95% confidence interval 0.31-0.91, P = .021 vs. low-volume surgeon, odds ratio 0.76, 95% confidence interval 0.69-0.82, P < .001). Among patients treated by male surgeons, a high rate of fragmented practice was associated with reduced odds of achieving a textbook outcome (male surgeons: odds ratio 0.76, 95% confidence interval 0.70-0.82, P < .001; female surgeons: odds ratio 0.81, 95% confidence interval 0.63-1.05, P = .110). Treatment by surgeons with higher fragmented practice was associated with higher expenditures (index expenditure: percentage difference 9.87, 95% confidence interval, 7.42-12.36; P < .05).

CONCLUSION

A high rate of fragmented practice adversely affected postoperative outcomes and healthcare expenditures even among high-volume surgeons with the impact varying based on surgeon sex.

Rescue from complications after pancreaticoduodenectomies at a low-volume Caribbean center: Value of tailored peri-pancreatectomy protocols

BACKGROUND

Pancreaticoduodenectomy (PD) is a technically complex operation, with a relatively high risk for complications. The ability to rescue patients from post-PD complications is as a recognized quality measure. Tailored protocols were instituted at our low volume facility in the year 2013.

AIM

To document the rate of rescue from post-PD complications with tailored protocols in place as a measure of quality.

METHODS

A retrospective audit was performed to collect data from patients who experienced major post-PD complications at a low volume pancreatic surgery unit in Trinidad and Tobago between January 1, 2013 and June 30, 2023. Standardized definitions from the International Study Group of Pancreatic Surgery were used to define post-PD complications, and the modified Clavien-Dindo classification was used to classify post-PD complications.

RESULTS

Over the study period, 113 patients at a mean age of 57.5 years (standard deviation [SD] ± 9.23; range: 30-90; median: 56) underwent PDs at this facility. Major complications were recorded in 33 (29.2%) patients at a mean age of 53.8 years (SD: ± 7.9). Twenty-nine (87.9%) patients who experienced major morbidity were salvaged after aggressive treatment of their complication. Four (3.5%) died from bleeding pseudoaneurysm (1), septic shock secondary to a bile leak (1), anastomotic leak (1), and myocardial infarction (1). There was a significantly greater salvage rate in patients with American Society of Anesthesiologists scores ≤ 2 (93.3% vs 25%; P = 0.0024).

CONCLUSION

This paper adds to the growing body of evidence that volume alone should not be used as a marker of quality for patients requiring PD. Despite low volumes at our facility, we demonstrated that 87.9% of patients were rescued from major complications. We attributed this to several factors including development of rescue protocols, the competence of the pancreatic surgery teams and continuous, and adaptive learning by the entire institution, culminating in the development of tailored peri-pancreatectomy protocols.

More problems, more money: Identifying and predicting high-cost rescue after colorectal surgery

Background

Successful rescue after elective surgery is associated with increased healthcare costs, but costs vary widely. Treating all rescue events the same may overlook targeted opportunities for improvement. The purpose of this study was to predict high-cost rescue after elective colorectal surgery.

Methods

We identified adult patients in the National Inpatient Sample (2016-2021) who underwent elective colectomy or proctectomy. Rescued patients were defined as those who underwent additional major procedures. Three groups were stratified: 1) uneventful recovery; 2) Low-cost rescue; 3) High-cost rescue. Multivariable Poisson regression was used to identify preoperative clinical predictors of high-cost versus low-cost rescue.

Results

We identified 448,590 elective surgeries, and rescued patients composed 4.8 %(21,635) of the total sample. The median increase in costs in rescued patients was $25,544(p < 0.001). Median total inpatient costs were $95,926 in the most expensive rescued versus $34,811 in the less expensive rescued versus $16,751 in the uneventfully discharged(p < 0.001). When comparing the secondary procedures between the less expensive and most expensive rescued groups, the most expensive had an increased proportion of reoperation (73.4 % versus 53.0 %,p < 0.001). When controlling for other factors and stratification by congestive heart failure due to an interaction effect, a reoperation was independently associated with high-cost rescue (RR with CHF = 3.29,95%CI:2.69-4.04; RR without CHF = 2.29,95%CI:1.97-2.67).

Conclusions

High-cost rescue after colorectal surgery is associated with disproportionately greater healthcare utilization and reoperation. For cost-conscious care, preemptive strategies that reduce reoperation-related complications can be prioritized.

Trends in Textbook Outcomes over Time: Are Optimal Outcomes Following Complex Gastrointestinal Surgery for Cancer Increasing?

BACKGROUND

The use of composite measures like "textbook outcome" (TO) may provide a more accurate measure of surgical quality. We sought to determine if TO has improved over time and to characterize the association of achieving a TO with trends in survival among patients undergoing complex gastrointestinal surgery for cancer.

METHODS

Medicare beneficiaries who underwent pancreas, liver, or colon resection for a cancer diagnosis between 2004 and 2016 were identified using the SEER-Medicare database. Rates of TO (no complication, extended length of stay, 90-day readmission, or 90-day mortality) were assessed over time.

RESULTS

Among 94,329 patients, 6765 (7.2%), 1985 (2.1%), and 85,579 (90.7%) patients underwent resection for primary pancreatic, hepatic, or colon cancer, respectively. In total, 53,464 (56.7%) patients achieved a TO; achievement of TO varied by procedure (pancreatectomy: 48.1% vs. hepatectomy: 55.2% vs. colectomy: 57.4%, p < 0.001). The proportion of patients achieving a textbook outcome increased over time for all patients (2004-2007, 53.3% vs. 2008-2011, 56.5% vs. 2012-2016, 60.1%) (5-year increase: OR 1.16 95%CI 1.13-1.18) (p < 0.001). Survival at 1-year following pancreatic, liver, or colon resection for cancer had improved over time among both patients who did and did not achieve a postoperative TO. TO was independently associated with a marked reduction in hazard of death (HR 0.44, 95%CI 0.43-0.45). The association of TO and survival was consistent among patients stratified by procedure.

CONCLUSION

Less than two-thirds of patients undergoing complex gastrointestinal surgery for a malignant indication achieved a TO. The likelihood of achieving a TO increased over time and was associated with improved survival.

Factors associated with failure to rescue after liver resection and impact on hospital variation: a nationwide population-based study

BACKGROUND

Failure to rescue (FTR) is defined as postoperative complications leading to mortality. This nationwide study aimed to assess factors associated with FTR and hospital variation in FTR after liver surgery.

METHODS

All patients who underwent liver resection between 2014 and 2017 in the Netherlands were included. FTR was defined as in-hospital or 30-day mortality after complications Dindo grade ≥3a. Variables associated with FTR and nationwide hospital variation were assessed using multivariable logistic regression.

RESULTS

Of 4961 patients included, 3707 (74.4%) underwent liver resection for colorectal liver metastases, 379 (7.6%) for other metastases, 526 (10.6%) for hepatocellular carcinoma and 349 (7.0%) for biliary cancer. Thirty-day major morbidity was 11.5%. Overall mortality was 2.3%. FTR was 19.1%. Age 65-80 (aOR: 2.86, CI:1.01-12.0, p = 0.049), ASA 3+ (aOR:2.59, CI: 1.66-4.02, p < 0.001), liver cirrhosis (aOR:4.15, CI:1.81-9.22, p < 0.001), biliary cancer (aOR:3.47, CI: 1.73-6.96, p < 0.001), and major resection (aOR:6.46, CI: 3.91-10.9, p < 0.001) were associated with FTR. Postoperative liver failure (aOR: 26.9, CI: 14.6-51.2, p < 0.001), cardiac (aOR: 2.62, CI: 1.27-5.29, p = 0.008) and thromboembolic complications (aOR: 2.49, CI: 1.16-5.22, p = 0.017) were associated with FTR. After case-mix correction, no hospital variation in FTR was observed.

CONCLUSION

FTR is influenced by patient demographics, disease and procedural burden. Prevention of postoperative liver failure, cardiac and thromboembolic complications could decrease FTR.

Impact of hospital quality on surgical outcomes in patients with high social vulnerability: Association of textbook outcomes and social vulnerability by hospital quality

BACKGROUND

We sought to define the impact of high- versus low-quality hospitals on the risk of adverse outcomes among patients undergoing hepatopancreatic surgery relative to social vulnerability. Social vulnerability is an important factor associated with risk of adverse postoperative outcomes.

METHODS

Patients from 2013 to 2017 were identified from the Medicare Inpatient Standard Analytic File. Hospital quality was determined by calculating risk-adjusted probability to achieve a textbook outcome. The Social Vulnerability Index was used to categorize patients. Risk-adjusted probability of mortality, morbidity, and textbook outcome was examined across varying social vulnerability indices stratified by low-, average-, and high-quality hospitals.

RESULTS

Among 27,000 patients who underwent a pancreatectomy (67%) or hepatectomy (33%%), median patient age was 72 years, 48% were female, and 89% were White; mean Social Vulnerability Index was 49. Risk-adjusted 90-day mortality (odds ratio: 1.32, 95% CI: 1.20-1.59, P = .004) and postoperative complications (odds ratio: 1.12, 95% confidence interval: 1.00-1.24, P = .044) were both higher among beneficiaries from the highest social vulnerability counties versus the lowest counties. At low-quality hospitals, patients from the highest vulnerability counties had 70% higher odds of mortality (odds ratio: 1.70, 95% confidence interval: 1.16-2.48, P = .007), 31% higher odds of overall morbidity odds ratio: 1.31, 95% confidence interval: 1.05-2.63, P = .013), and 19% lower odds of achieving a textbook outcome (odds ratio: 0.81, 95% confidence interval: 0.66-0.99, P = .035)-all of which were markedly worse compared with outcomes achieved at high-quality hospitals.

CONCLUSION

Among patients with increased social vulnerability, outcomes were considerably better at high-quality hospitals. Referral of socially vulnerable patients to high-quality hospitals represents an important opportunity to ensure optimal outcomes after complex surgery.

Prehabilitation in hepato-pancreato-biliary surgery: A systematic review and meta-analysis. A necessary step forward evidence-based sample size calculation for future trials

INTRODUCTION

Prehabilitation is defined as preoperative conditioning of patients in order to improve post-operative outcomes. Some studies showed an increase in functional recovery following colorectal surgery, but its effect in hepato-pancreato-biliary (HPB) surgery is unclear. The aim of this study was to realize a systematic literature review and meta-analysis on the current available evidence on prehabilitation in HPB surgery.

MATERIALS AND METHODS

A systematic review and a metanalysis were carried out on prehabilitation (physical, nutritional and psychological interventions) in HPB surgery (2009-2019). Assessed outcomes were postoperative complications, length of stay (LOS), 30-day readmission, and mortality.

MAIN RESULTS

Four studies among the 191 screened were included in this systematic review (3 randomized controlled trials, 1 case-control propensity score study), involving 419 patients (prehabilitation group, n=139; control group, n=280). After pooling, no difference was observed on LOS ((-4.37 days [95% CI: -8.86; 0.13]) or postoperative complications (RR 0.83 [95%CI: 0.62; 1.10]), reported by all the included studies. Two trials reported on readmission rate, but given the high heterogeneity, a meta-analysis was not realized. No deaths were reported among the included studies.

CONCLUSION

No effect of prehabilitation programs in HPB surgery was observed on LOS or postoperative complications rate. Future trials with standardized outcomes of measure, and adequately powered samples calculations are thus required.

PROSPERO REGISTRATION

CRD42020165218.

Training Paradigms in Hepato-Pancreatico-Biliary Surgery: an Overview of the Different Fellowship Pathways

Hepato-pancreatico-biliary (HPB) surgery, and the training of HPB surgeons, has evolved significantly over the last several decades. The current state of training in HPB surgery in North America is defined through three main pathways: the Complex General Surgical Oncology (CGSO) ACGME fellowship, the American Society of Transplant Surgeons (ASTS) fellowship, and the Americas Hepatopancreaticobiliary Association (AHPBA) fellowship. These fellowships offer variable experiences in pancreas, liver, and biliary cases, and each pathway offers a unique perspective on HPB surgery. The CGSO ACGME, ASTS, and AHPBA fellowships represent decades of work by the three major surgical leadership stakeholders to improve and ensure quality training of future HPB surgeons. The best care is provided by the HPB surgeon who has been trained to understand the importance of all available treatment options within the context of a multidisciplinary setting. The three fellowship pathways are outlined in this paper with the nuances and variations characteristic of the different training programs highlighted.

High Social Vulnerability and "Textbook Outcomes" after Cancer Operation

BACKGROUND

The effect of community-level factors on surgical outcomes has not been well examined. We sought to characterize differences in "textbook outcomes" (TO) relative to social vulnerability among Medicare beneficiaries who underwent operations for cancer.

METHODS

Individuals who underwent operations for lung, esophageal, colon, or rectal cancer between 2013 and 2017 were identified using the Medicare database, which was merged with the CDC's Social Vulnerability Index (SVI). TO was defined as surgical episodes with the absence of complications, extended length of stay, readmission, and mortality. The association of SVI and TO was assessed using mixed-effects logistic regression.

RESULTS

Among 203,800 patients (colon, n = 113,929; lung, n = 70,642; rectal, n = 14,849; and esophageal, n = 4,380), median age was 75 years (interquartile range 70 to 80 years) and the overwhelming majority of patients was White (n = 184,989 [90.8%]). The overall incidence of TO was 56.1% (n = 114,393). The incidence of complications (low SVI: 21.5% vs high SVI: 24.0%) and 90-day mortality (low SVI: 7.0% vs high SVI: 8.4%) were higher among patients from highly vulnerable neighborhoods (both, p < 0.05). In turn, there were lower odds of achieving TO among high-vs low-SVI patients (odds ratio 0.83; 95% CI, 0.78 to 0.87). Although high-SVI White patients had 10% lower odds (95% CI, 0.87 to 0.93) of achieving TO, high-SVI non-White patients were at 22% lower odds (95% CI, 0.71 to 0.85) of postoperative TO. Compared with low-SVI White patients, high-SVI minority patients had 47% increased odds of an extended length of stay, 40% increased odds of a complication, and 23% increased odds of 90-day mortality (all, p < 0.05).

CONCLUSIONS

Only roughly one-half of Medicare beneficiaries achieved the composite optimal TO quality metric. Social vulnerability was associated with lower attainment of TO and an increased risk of adverse postoperative surgical outcomes after several common oncologic procedures. The effect of high SVI was most pronounced among minority patients.

County-Level Social Vulnerability is Associated with Worse Surgical Outcomes Especially Among Minority Patients

OBJECTIVE

We sought to characterize the association between patient county-level vulnerability with postoperative outcomes.

SUMMARY BACKGROUND DATA

While the impact of demographic-, clinical- and hospital-level factors on outcomes following surgery have been examined, little is known about the effect of a patient's community of residence on surgical outcomes.

METHODS

Individuals who underwent colon resection, coronary artery bypass graft (CABG), lung resection, or lower extremity joint replacement (LEJR) were identified in the 2016-2017 Medicare database, which was merged with CDC vulnerability index (SVI) dataset at the beneficiary level of residence. Logistic regression models were utilized to estimate the probability of postoperative complications, mortality, readmission, and expenditures.

RESULTS

Among 299,583 Medicare beneficiary beneficiaries who underwent a colectomy (n = 88,778, 29.6%), CABG (n = 109,564, 36.6%), lung resection (n = 30,401, 10.1%), or LEJR (n = 70,840, 23.6%).Mean SVI score was 50.2 (SD: (25.2); minority patients were more likely to reside in highly vulnerable communities (low SVI: n = 3,531, 5.8% vs. high SVI: n = 7,895, 13.3%; p < 0.001). After controlling for competing risk factors, the risk-adjusted probability of a serious complication among patients from a high versus low SVI county was 10-20% higher following colectomy (OR 1.1 95%CI 1.1-1.2) or CABG (OR 1.2 95%CI 1.1-1.3), yet there no association of SVI with risk of serious complications following lung resection (OR 1.2 95%CI 1.0-1.3) or LEJR (OR 1.0 95%CI 0.93-1.2). The risk-adjusted probability of 30-day mortality was incrementally higher among patients from high SVI counties following colectomy (OR 1.1 95%CI 1.1-1.3), CABG (OR 1.4, 95%CI 1.2-1.5), and lung resection (OR 1.4 (95%CI 1.1-1.8), yet not LEJR (OR 0.95 95%CI 0.72-1.2). Black/minority patients undergoing a colectomy, CABG, or lung resection who lived in highly socially vulnerable counties had an estimate 28-68% increased odds of a serious complication and a 58-60% increased odds of 30-day mortality compared with a black/minority patient from a low socially vulnerable county, as well as a markedly higher risk than white patients (all p > 0.05).

CONCLUSION

Patients residing in vulnerable communities characterized by a high SVI generally had worse postoperative outcomes. The impact of social vulnerability was most pronounced among black/minority patients, rather than white individuals. Efforts to ensure equitable surgical outcomes need to focus on both patient-level, as well as community-specific factors.

Quality Versus Costs Related to Gastrointestinal Surgery: Disentangling the Value Proposition

BACKGROUND

There has been a dramatic increase in worldwide health care spending over the last several decades. Operative procedures and perioperative care in the USA represent some of the most expensive episodes per patient. In view of both the rising cost of health care in general and the rising cost of surgical care specifically, policymakers and stakeholders have sought to identify ways to increase the value-improving quality of care while controlling (or diminishing) costs. In this context, we reviewed data relative to achieving the "value proposition" in the delivery of gastrointestinal surgical care.

METHODS

The National Library of Medicine online repository (PubMed) was text searched for human studies including "cost," "quality," "outcomes," "health care," "surgery," and "value." Results from this literature framed by the Donabedian conceptual model (identifying structures, processes, and outcomes), and the resulting impact of efforts to improve quality on costs.

RESULTS

The relationship between quality and costs was nuanced. Better quality care, though associated with better outcomes, was not always reported as concomitant with low costs. Moreover, some centers reported higher costs of surgical care commensurate with higher quality. Conversely, higher costs in health care delivery were not always linked to improved outcomes. While higher quality surgical care can lead to lower costs, higher costs of care were not necessarily associated with better outcomes. Strategies to improve quality, reduce cost, or achieve both simultaneously included regionalization of complex operations to high-volume centers of excellence, overall reduction in complications, introducing evidence-based improvements in perioperative care pathways including as enhanced recovery after surgery (ERAS), and elimination of inefficient or low-value care.

CONCLUSIONS

The relationship between quality and cost following gastrointestinal surgical procedure is complex. Data from the current study should serve to highlight the various means available to improve the value proposition related to surgery, as well as encourage surgeons to become more engaged in the national conversation around the Triple Aim of better health care quality, lower costs, and improved health care outcomes.

Interaction of Surgeon Volume and Nurse-to-Patient Ratio on Post-operative Outcomes of Medicare Beneficiaries Following Pancreaticoduodenectomy

BACKGROUND

We sought to assess the effect of nurse-to-patient ratio on outcomes with a focus on defining whether nurse-to-patient ratio altered outcomes relative to pancreaticoduodenectomy (PD) surgeon specific volume.

METHODS

Medicare SAFs from 2013-2015 were used to identify patients who underwent PD. Nurse-to-patient ratio, PD specific surgeon volume were stratified. Association of factors associated with short term outcomes was evaluated.

RESULTS

Overall, 6668 patients (median age 73, IQR 68-77; 52.8% male) were identified. The median annual PD volume of surgeons in the highest volume tier was 24 (IQR 21-29), whereas surgeons in the lowest tier performed 2 PDs annually (IQR 1-3) (p < 0.001). Compared with hospitals that had the highest nurse-to-patient ratio tier, patients at hospitals with the lowest nurse-to-patient ratio tier were 26% more likely to have a complication (OR 1.26, 95% CI 1.02-1.55). Additionally, patients of surgeons in the lowest tier had 43% greater odds of suffering a complication compared to patients of surgeons in the highest tier (OR 1.43, 95% CI 1.11-1.84). However, patients who underwent a PD by a surgeon within the lowest tier had similar odds of a complication irrespective of nurse-to-patient ratio (OR 1.34, 95% CI 0.97-1.86).

CONCLUSION

Compared with patients who underwent an operation by a surgeon in highest PD volume tier, patients treated by surgeons in the lowest tier had higher odds of post-operative complications which was not mitigated by a higher nurse-to-patient ratio.

Routine Intensive Care Unit Admission Following Liver Resection: What Is the Value Proposition?

INTRODUCTION

The value of routine ICU admission after elective surgery has been debated due to the lack of robust evidence supporting its benefit, as well as the increased incurred costs. We sought to analyze outcomes of patients undergoing hepatectomy who were routinely admitted to the intensive care unit (ICU) compared with surgical ward admission.

METHODS

Patients were identified in the Truven Health Analytics MarketScan Commercial Claims and Encounters Database from 2010 to 2016. Routine postoperative ICU admission was defined as ICU admission for ≤ 24 h on postoperative day 0. Potential association between routine ICU admission with extended length-of-stay (LOS), failure-to-rescue, and total inpatient costs was analyzed.

RESULTS

In total 7970 patients underwent hepatectomy; 37.7% (n = 3001) had routine ICU admission and 62.3% (n = 4969) surgical ward admission. Among the 3001 patients who had routine ICU admission, 1137 (37.9%) had a major and 1864 (62.1%) had a minor hepatectomy. Routine ICU admission was not associated with lower failure-to-rescue (routine ICU 4.9% vs. ward 1.8%; p < 0.001). Patients routinely admitted to the ICU had longer median LOS (routine ICU 7 days, IQR 5-15 days vs. ward 5 days, IQR 4-7 days; p < 0.001). Median payments were higher for routine ICU admission than for surgical ward admission ($50,501, IQR $34,270-$80,459 vs. $39,774, IQR $28,555-$58,270, respectively).

CONCLUSION

Routine ICU admission was associated with longer LOS and higher hospital payments, yet did not translate into lower failure-to-rescue among patients undergoing hepatectomy.

Use of Machine Learning for Prediction of Patient Risk of Postoperative Complications After Liver, Pancreatic, and Colorectal Surgery

BACKGROUND

Surgical resection is the only potentially curative treatment for patients with colorectal, liver, and pancreatic cancers. Although these procedures are performed with low mortality, rates of complications remain relatively high following hepatopancreatic and colorectal surgery.

METHODS

The American College of Surgeons (ACS) National Surgical Quality Improvement Program was utilized to identify patients undergoing liver, pancreatic and colorectal surgery from 2014 to 2016. Decision tree models were utilized to predict the occurrence of any complication, as well as specific complications. To assess the variability of the performance of the classification trees, bootstrapping was performed on 50% of the sample.

RESULTS

Algorithms were derived from a total of 15,657 patients who met inclusion criteria. The algorithm had a good predictive ability for the occurrence of any complication, with a C-statistic of 0.74, outperforming the ASA (C-statistic 0.58) and ACS-Surgical Risk Calculator (C-statistic 0.71). The algorithm was able to predict with high accuracy thirteen out of the seventeen complications analyzed. The best performance was in the prediction of stroke (C-statistic 0.98), followed by wound dehiscence, cardiac arrest, and progressive renal failure (all C-statistic 0.96). The algorithm had a good predictive ability for superficial SSI (C-statistic 0.76), organ space SSI (C-statistic 0.76), sepsis (C-statistic 0.79), and bleeding requiring transfusion (C-statistic 0.79).

CONCLUSION

Machine learning was used to develop an algorithm that accurately predicted patient risk of developing complications following liver, pancreatic, or colorectal surgery. The algorithm had very good predictive ability to predict specific complications and demonstrated superiority over other established methods.

Novel Machine Learning Approach to Identify Preoperative Risk Factors Associated With Super-Utilization of Medicare Expenditure Following Surgery

Importance

Typically defined as the top 5% of health care users, super-utilizers are responsible for an estimated 40% to 55% of all health care costs. Little is known about which factors may be associated with increased risk of long-term postoperative super-utilization.

Objective

To identify clusters of patients with distinct constellations of clinical and comorbid patterns who may be associated with an elevated risk of super-utilization in the year following elective surgery.

Design, Setting, and Participants

A retrospective longitudinal cohort study of 1 049 160 patients who underwent abdominal aortic aneurysm repair, coronary artery bypass graft, colectomy, total hip arthroplasty, total knee arthroplasty, or lung resection were identified from the 100% Medicare inpatient and outpatient Standard Analytic Files at all inpatient facilities performing 1 or more of the evaluated surgical procedures from 2013 to 2015. Data from 2012 to 2016 were used to evaluate expenditures in the year preceding and following surgery. Using a machine learning approach known as Logic Forest, comorbidities and interactions of comorbidities that put patients at an increased chance of becoming a super-utilizer were identified. All comorbidities, as defined by the Charlson (range, 0-24) and Elixhauser (range, 0-29) comorbidity indices, were used in the analysis. Higher scores indicated higher comorbidity burden. Data analysis was completed on November 16, 2018.

Main Outcome and Measures

Super-utilization of health care in the year following surgery.

Results

In total, 1 049 160 patients met inclusion criteria and were included in the analytic cohort. Their median (interquartile range) age was 73 (69-78) years, and approximately 40% were male. Super-utilizers comprised 4.8% of the overall cohort (n = 79 746) yet incurred 31.7% of the expenditures. Although the difference in overall expenditures per person between super-utilizers ($4049) and low users ($2148) was relatively modest prior to surgery, the difference in expenditures between super-utilizers ($79 698) vs low users ($2977) was marked in the year following surgery. Risk factors associated with super-utilization of health care included hemiplegia/paraplegia (odds ratio, 5.2; 95% CI, 4.4-6.2), weight loss (odds ratio, 3.5; 95% CI, 2.9-4.2), and congestive heart failure with chronic kidney disease stages I to IV (odds ratio, 3.4; 95% CI, 3.0-3.9).

Conclusions and Relevance

Super-utilizers comprised only a small fraction of the surgical population yet were responsible for a disproportionate amount of Medicare expenditure. Certain subpopulations were associated with super-utilization of health care following surgical intervention despite having lower overall use in the preoperative period.

Insurance Coverage Type Impacts Hospitalization Patterns Among Patients with Hepatopancreatic Malignancies

INTRODUCTION

Disparities in health and healthcare access remain a major problem in the USA. The current study sought to investigate the relationship between patient insurance status and hospital selection for surgical care.

METHODS

Patients who underwent liver or pancreatic resection for cancer between 2004 and 2014 were identified in the National Inpatient Sample. The association of insurance status and hospital type was examined.

RESULTS

In total, 22,254 patients were included in the study. Compared with patients with private insurance, Medicaid patients were less likely to undergo surgery at urban non-teaching hospitals (OR = 0.36, 95%CI 0.22-0.59) and urban teaching hospitals (OR = 0.54, 95%CI 0.34-0.84) than rural hospitals. Medicaid patients were less likely to undergo surgery at private investor-owned hospitals (OR = 0.53, 95%CI 0.38-0.73) than private non-profit hospitals. In contrast, uninsured patients were 2.2-fold more likely to go to government-funded hospitals rather than private non-profit hospitals (OR = 2.19, 95%CI 1.76-2.71).

CONCLUSION

Insurance status was strongly associated with the type of hospital in which patients underwent surgery for liver and pancreatic cancers. Addressing the reasons for inequitable access to different hospital settings relative to insurance status is essential to ensure that all patients undergoing pancreatic or liver surgery receive high-quality surgical care.

In-hospital Mortality Following Pancreatoduodenectomy: a Comprehensive Analysis

BACKGROUND

While patient- and hospital-level factors affecting outcomes of patients undergoing pancreatoduodenectomy (PD) have been well described separately, the relative impact of these factors on in-hospital mortality has not been comprehensively assessed.

METHODS

Retrospective review of the National Inpatient Sample database (January 2004-December 2014) was conducted to identify patients undergoing PD. Factors associated with in-hospital mortality after PD were analyzed after adjusting for previously defined patient- and hospital-level risk factors.

RESULTS

A total of 9639 patients who underwent a PD at 2325 hospitals were identified. Median patient age was 57 years (IQR 66-73). Overall, mortality following PD was 3.2%. When patient- and hospital-level characteristics were analyzed in the same model, patient-level characteristic associated with increased odds of in-hospital mortality included increasing patient age (OR 1.05, 95% CI 1.03-1.06/per 5 years increase), male sex (OR 1.47, 95% CI 1.16-1.86), the presence of liver disease (OR 3.03, 95% CI 1.99-4.61), chronic kidney disease (OR 1.78, 95% CI 1.18-2.68), and congestive heart failure (OR 2.48, 95% CI 1.65-3.74). The only hospital characteristic associated with odds of mortality following PD included compliance with Leapfrog volume standards (OR 0.70, 95% CI 0.54-0.92).

CONCLUSION

Patient-level factors, such as advanced comorbidities, male sex, and increased age, contributed the most to increased risk of mortality after PD. Hospital volume was the only hospital-level factor contributing to risk of in-hospital mortality following PD.