69.09 Timely Access to Care for Patients with Traumatic Brain Injuries

Posted on January 18, 2018

B. Tracy1, M. E. Barnett1, C. Spencer1, A. Butcher1, M. Brown1, Z. Stombaugh1, J. Dunne1  1Memorial University Medical Center-Mercer University School Of Medicine,Surgery,Savannah, GEORGIA, USA

Introduction:  Recent studies support direct transport of patients sustaining TBI to Level 1 trauma centers (L1TC) due to decreased morbidity and mortality.  However, in some rural areas, TBI patients are triaged to the closest non-trauma center (NTC), then transferred to a L1TC.  Therefore, we sought to determine the effectiveness of our rural trauma system by comparing the outcomes of TBI patients initially triaged to a NTC compared to those transported directly to a L1TC.

Methods:  Prospective data were collected on 8,558 patients admitted to a rural L1TC over a 3- year period from 2012–2015. TBI was identified in 1,719 patients and comprised the study cohort.  Patients were stratified by age, GCS, ISS, and transfer status.  Outcome variables included # of TBI interventions, hospital length of stay (HLOS), ICU length of stay (ICU LOS) and mortality. In addition, multivariate regression was used to determine risk factors for time to transfer, HLOS and mortality. 

Results: The mean age of the study cohort was 51±21, mean ISS was 15±9, mean GCS was 13±4, with 49.7% of patients transferred from other facilities.  The most common mechanisms of injury included MVC (45%), falls (37%) and other blunt trauma (18%).  Overall mortality was 7%.  Patients transported directly to a L1TC were significantly younger (47±20 vs 56±23, p < 0.001), more severely injured based on ISS (14±8 vs 12±7, p < 0.001) and had increased mortality (8% vs 5%, p < 0.01) compared to patients initially triaged to a NTC.  The study cohort was then stratified into 3 categories based on GCS to further delineate these differences: severe TBI (GCS 3-8), moderate TBI (GCS 9-12) and mild TBI (GCS 13-15).  Patients with severe TBI were significantly younger, more severely injured, more often transported directly to a L1TC and had a higher mortality when compared to patients with mild and moderate TBI (p< 0.01).  Severe TBI patients also had significantly longer HLOS and ICU LOS and more TBI interventions compared to patients with mild and moderate TBI (p < 0.001).  Multivariate regression revealed decreasing age, decreasing GCS and increasing ISS as risk factors for direct transport to L1TC and decreasing GCS, increasing ISS and decreased transport time as risk factors for increased HLOS and mortality.

Conclusion: Patients transported directly to a L1TC were significantly younger, more severely injured and had increased mortality when compared to patients initially triaged to a NTC. Risk factors for transport directly to a L1TC include decreasing age, worsening neurologic status and increasing injury severity.  Finally, worsening neurologic status, increasing injury severity and shorter transport times were risk factors for increased HLOS and mortality.  In summary, the most critically injured patients were transported to the L1TC directly indicating that the right patients were being transported to the proper facility at the appropriate time.
 

69.10 Scholarly Activity in Academic Plastic Surgery: The Gender Difference

Posted on January 18, 2018

S. E. Sasor1, J. A. Cook1, S. P. Duquette1, T. A. Evans1, S. S. Tholpady1, M. W. Chu1, L. G. Koniaris1  1Indiana University,Plastic Surgery,Indianapolis, IN, USA

Purpose:

The number of women in medicine has grown rapidly in recent years. Women constitute over 50% of medical school graduates and hold 38% of faculty positions at U.S. medical schools. Despite this, gender disparities remain prevalent in most surgical subspecialties, including plastic surgery. The purpose of this study is to analyze female authorship trends to identify factors that affect scholarly output and professional advancement in academic plastic surgery.

 

Methods:

A cross-sectional study of all academic plastic surgeons was performed. Data was collected on gender, degrees, titles, and affiliation residency programs from department websites and online resources. National Institute of Health (NIH) funding was determined using the Research Portfolio Online Reporting Tools database. Number of published manuscripts, citations, and h-index were obtained from Scopus (Elsevier Inc., New York, NY). Statistical analysis was performed in SPSS.

 

Results:

A total of 814 plastic surgeons were identified from 91 training programs in the US. Mean years in practice was 15.3 and the majority were male (83.2%). Average number of published manuscripts, citations, and h-index were 45.5, 974.9, and 11.6, respectively.

 

Compared to male surgeons, female surgeons had significantly fewer years in practice (9.9 vs. 16.4, p<0.001), held lower academic ranks (57.3% vs. 33.9% assistant professor, p<0.001 compared to 6.6% vs. 29.0% full professor, p<0.001), and published less (20.0 vs. 50.2 manuscripts, p<0.001, and 7.3 vs 12.6 h-index, p<0.001). Females with titles of assistant or associate professor had significantly fewer years in practice than males with the same title (6.1 vs. 9.0 years, p<0.001 assistant and 13.5 vs. 15.8 years p=0.01 associate). The trend continued for full professors but was not significant. Males published more manuscripts, had more citations and higher h-indexes than female surgeons at each academic rank but the findings were not statistically significant beyond the level of assistant professor.

 

Seven of 136 female surgeons (5.1%) and 46 of 678 male surgeons (6.9%) received an NIH grant during their career (p=0.57). Mean number of awards and total amount of funding was 10.3/$2,996,734 for women and 8.0/$1,853,345 for men (p=0.469 and p=0.57, respectively). NIH-funded surgeons of both sexes published more than non-funded surgeons (119.5 vs. 40.3 publications, p<0.001 and 24.7 vs. 10.7 h-index, p<0.001). There was no gender difference in scholarly output amongst NIH-funded surgeons.

 

Conclusion:

Research productivity is a metric for promotion in academic medicine. This study identifies significant gender disparities in scholarly productivity amongst plastic surgeons in academia. Differences are most apparent at junior ranks, suggesting that women who remain in academic medicine may later overcome publishing barriers faced earlier in their careers. 

 

69.07 Regional Variation in Laparoscopic and Open Inguinal Hernia Repair Across Michigan

Posted on January 18, 2018

J. V. Vu1, V. Gunaseelan1,2, E. Seese2, M. J. Englesbe1,2, G. L. Krapohl1,2, D. A. Campbell1,2, D. Telem1  2Michigan Surgical Quality Collaborative,Ann Arbor, MICHIGAN, USA 1University Of Michigan,Department Of Surgery,Ann Arbor, MI, USA

Background: Despite the benefits of minimally invasive surgery (MIS) for inguinal hernia repair, adoption of this technique has been suboptimal compared to other operations (e.g., colectomy, hysterectomy, cholecystectomy). To better understand the clinical, demographic, and provider factors associated with uptake of MIS for inguinal hernia, we characterized the variation in utilization rate across a statewide population.

Methods: A retrospective cohort study was performed in patients undergoing open and MIS inguinal hernia repair from 2012 to 2016, using data from the Michigan Surgical Quality Collaborative, a statewide surgical collaborative of 73 hospitals. Operations for recurrent hernia, strangulation, or gangrene were excluded. Primary outcome was MIS utilization rate by Dartmouth hospital referral region. Secondary outcomes were provider, patient, and hospital characteristics associated with MIS utilization. Data were evaluated using a weighted hierarchical logistic regression model.

Results: A total of 6,185 inguinal hernia repairs were identified. Utilization of MIS ranged from 13.5% to 42.8% of all repairs across six geographic regions, as demonstrated in Figure 1 (p<0.001). Hospital site accounted for 41% of the variation in utilization; MIS repair was associated with higher hospital volume (p<0.001), but not with teaching status. Out of the 558 total surgeons, 322 (58%) performed no MIS repairs. After controlling for clinical, demographic, and geographic factors, patients aged 18 – 44 (OR 1.68, p<0.001) and patients aged 45 – 64 (OR 1.49, p<0.001) were more likely to receive MIS than patients aged 65 and older. Black patients were less likely to receive MIS than white patients (OR 0.72, p=0.017). Patients with COPD, hypertension, and in American Society of Anesthesiologists Class III or IV were also significantly less likely to undergo MIS repair. There were no differences in MIS utilization with respect to gender, obesity, and tobacco use.

Conclusion: Utilization of MIS for inguinal hernia repair is widely variable by region across a statewide population. Causes of this variation are likely multifactorial and are attributable to hospital, provider, and patient factors. Over half of the surgeons sampled do not perform MIS repair. Additionally, variation in MIS utilization was independently associated with patient age, race, and comorbid conditions. These findings support the presence of a practice gap in the delivery of MIS care. Exploratory, in-depth qualitative work investigating provider and patient-level barriers to MIS inguinal hernia repair is needed to develop evidence-based implementation intervention strategies.

69.08 Firearm Violence in Miami: Through The Lens of a Trauma Care Provider

Posted on January 18, 2018

A. Sarver1, K. Polcari1, K. Brownlee1, R. Rattan1, T. Zakrison1  1University Of Miami,Miami, FL, USA

Introduction:
Large gaps of knowledge remain in the understanding of social complications experienced by survivors of firearm violence and interventions that could be employed to reduce its incidence. Little is known about the perceptions of trauma care providers about their understanding of a firearm violence survivor’s experience – especially at  different stages of trauma care. This is important as trauma providers frequently inform the creation of firearm-related injury prevention programs. The objective of this study is to understand these perceptions of trauma providers about the causes of firearm violence in our community of South Florida and furthermore assess their thoughts regarding the experience of firearm violence for their gunshot wound patients. 

Methods:
We used a mixed methods approach to triangulate both qualitative and quantitative results. Qualitatively, we performed individual interviews with semi-structured, open ended questions. Participants were trauma surgeons and trauma anesthesiologists working at a high volume, Level 1 trauma center that sees a significant rate of firearm-related violence. Participants were identified using a snowballing sampling technique. Interviews continued until thematic saturation occurred. Grounded theory methodology was employed to build theory from salient themes. Observational data was incorporated from interviews with participants. Quantitative data included similar, survey-type questions with results quantified using Likert scales. 

Results:
A total of 12 trauma providers were interviewed over a three month period. The immediate experience of gun violence was interpreted as having a fear of dying, confusion and desire for revenge. The initial resuscitation of these patients was felt to be dehumanizing at times while the impact of injury was devastating to families, emotionally and economically. Young low-income Black males are disproportionately affected in Miami, yet gun violence was reiterated to affect all. Socio-demographic inequality was identified by all providers as a major reason for violence with gun control having mixed responses. Hospital-based injury prevention programs are supported by providers, with leadership of such programs emanating from the communities. Overall observations showed that many of the providers spoke from a place of detachment, and cultural incompetence. Many of the ideas  reflected a misunderstanding or unwillingness to recognize gun violence as a complex system of targeted oppression. Quantitative analysis demonstrated similar findings.

Conclusion:
Firearm violence is identified by most trauma care providers as rooted in social inequality. The role of gun control is mixed. While the experience of being injured by firearm violence is identified as a deeply frightening for both patient and family, there is a disconnect between the perceptions of trauma care providers and possibly with patients. This merits further exploration for formal comparison.  
 

69.05 Disparities in Esophageal Cancer: The Role of Insurance In time To Surgery and Outcomes.

Posted on January 18, 2018

J. D. Borgella1, F. Espinoza-Mercado1, T. Imai1, R. Alban1, H. Soukiasian1  1Cedars-Sinai Medical Center,Department Of Thoracic Surgery,Los Angeles, CALIFORNIA, USA

Introduction: Previous studies have shown disparities in esophageal cancer-related mortality rates among different races and those with lower socioeconomic status. Very little data exists on the role of insurance in outcomes of esophageal cancer, particularly for surgical candidates. We sought to identify the differences in the uninsured vs. insured patient population including the time interval from diagnosis to surgery (DTS) and survival.

Methods: The National Cancer Data Base (NCDB) was queried for patients diagnosed with pathologic Stage 0-IV esophageal cancer from 2004-2014. Insurance status was stratified into two groups: Insured vs. Uninsured.  Demographics and clinical characteristics were analyzed between groups. Time interval for DTS was estimated for both groups; overall survival was calculated using the Kaplan-Meier curves and the log-rank test was used to compare survival distributions.

Results: From 2004-2014 there were 125,493 patients diagnosed with esophageal cancer. Of which 4,495 (3.6%) were uninsured. Of these, 787 (17.5%) received a surgical procedure compared to 31.5% of the 120,998 insured patients. 42.2% of uninsured patients were less than 55 years old vs. 14.5% of insured patients. Blacks and Hispanics represented 23.2% and 8.8% of uninsured patients vs. 9.2% and 3% of insured patients. 31.9% of uninsured patients made less than $38K a year vs. 18% of insured patients. Interestingly uninsured patients had a lower Charlson-Deyo score compared to insured patients (CD = 0, 79.7% vs. 72.5%, respectively p< 0.001). Of all patients who received a definitive surgical procedure, uninsured patients waited longer for surgery (105 days vs. 94 days, p<0.01). The greatest differences were noted in Stage I disease, where the difference in DTS was 87 days, interquartile range (IQR) 64-119 days compared to insured patients (68 days, IQR 27-98 days, p=0.034). Uninsured patients had a statistically significant worse overall survival at every stage, particularly in stage I which had a survival of 26.3 months for the uninsured compared to 46.9 months in the insured (p< 0.01).  There was no statistically significant difference in the two groups from time of diagnosis to chemotherapy or radiation therapy.

Conclusion: Uninsured patients with stage 1 esophageal cancer receive definitive surgical treatment later than their insured counterparts and have worse survival. This disparity is also noted on other stages. Understanding the role of insurance in access to care for this cohort warrants further analysis.

 

69.06 Race Does Not Affect Length of Stay in Colorectal ERAS Patients with Post-Operative Complications.

Posted on January 18, 2018

P. K. Patel1, D. I. Chu1, L. Goss1, J. G. Wiener1, T. S. Wahl1, K. D. Cofer1, J. S. Richman1, M. S. Morris1, J. A. Cannon1, G. D. Kennedy1  1University Of Alabama at Birmingham,Gastrointestinal Surgery,Birmingham, Alabama, USA

Introduction:
Racial disparities have been documented in surgical outcomes. For example, recent studies have shown that black patients have longer length-of-stays (LOS) and higher rates of post-operative complications (POCs) than similar white patients.  However, it is unclear if disparities persist between black and white patients who suffer similar inpatient POCs. The primary aim of this study was to assess differences between minority status (white vs. non-white) in the LOS of patients who suffered complications for surgical patients with and without the Enhanced Recovery After Surgery (ERAS) pathway. We hypothesized that minority patients will have longer LOS if they suffer POCs.

Methods:
Using a prospectively maintained database of patients undergoing colorectal surgery before and after the implementation of ERAS at a single institution, we identified patients who suffered a POC using NSQIP variables. The primary outcome was LOS. Bivariate comparisons were made between races and ERAS status using chi square tests, one-way analysis of variances, and the Kruskal-Wallis test.

Results:
Out of a total of 1121 patients, 718 (64%) were pre-ERAS and 403 (36%) underwent ERAS. Of the 718 pre-ERAS patients, 191 (26.6%) were members of a minority group and 527 (73.4%) were white.  A total of 201 (28.0%) patients had post-operative complications with rates of 36.1% for minorities and 25.1% for whites (p<0.01). Overall there was a significant difference in LOS between minorities and whites (median 7 days vs. 5 days, p<0.01). Among those without POCs, minorities had a longer LOS (median 6 days vs. 5 days, p=0.01), but there was no difference in LOS for those with a POC (median 8 days vs. 8 days, p=0.95). Of the 403 ERAS patients, 221 (28.9%) were members of minority groups and 543 (71.1%) were white. A total of 106 (26.3%) individuals experienced a POC with rates of 28.6% for minorities and 25.4% for whites (p=0.50). Top complications in ERAS patients were organ space SSI (4.7%), superficial incisional SSI (3.4%), and UTI (2.5%). There was no significant difference in POCs by minority status (p=0.93). LOS did not differ by minority status overall (minority median 3 days vs. 3 days, p=0.50), for patients without complications (3 days vs. 3 days, p=0.58), or for patients with POCs (minority median 6 days vs. 5 days, p=0.26).

Conclusion:
In the pre-ERAS era, racial disparities existed with longer LOS and higher rates of POCs among minority patients. These disparities in LOS, however, appeared to be driven by patients without POCs. Under ERAS, there were no observed racial disparities in LOS, with or without POCs. The effect of ERAS on reducing disparities in LOS may therefore occur through its standardization of recovery pathways for patients without complications.

69.04 Surgery on Sunday Louisville: Achieving the Quadruple Aim by providing free surgery for the uninsured

Posted on January 18, 2018

S. C. Walling2, J. C. Heimroth1,2, E. R. Sutton1,2  1University Of Louisville,Department Of Surgery,Louisville, KY, USA 2Surgery On Sunday Louisville, Inc,Louisville, KY, USA

Introduction: In a climate of increasing expectation and perpetual change, every healthcare organization struggles to balance its efforts to improve patient experience and outcomes while reducing costs and fostering the well-being of its personnel.  The pursuit of these four objectives—the Quadruple Aim—has become a topic of national interest.  Surgery on Sunday Louisville (SOSL), a non-profit organization in Louisville, KY, sought to achieve this target by partnering with the local medical community to provide free surgery to the underinsured.  The objective of this study is to show that SOSL serves as a model of meeting the Quadruple Aim by improving population health, enhancing the patient experience, reducing the cost of care, and supporting provider well-being.

Methods: A retrospective analysis of SOSL’s outcomes and practices was performed in order to evaluate its efficacy in achieving the 4 goals of the Quadruple Aim.  Its effect on population health and patient experience were assessed by analyzing data regarding the utilization of its services and the characteristics of its patients and their medical problems.  A survey administered to a cohort of 25 volunteers was utilized to evaluate the effect of the program on the resilience and well-being of SOSL’s volunteer staff.

Results: Now in its fifth year, SOSL has improved population health by caring for 363 patients in its outpatient surgical clinic and free surgical events and performing 81 surgeries and 144 endoscopies, which have not only reduced the burden of disease but also increased access to colorectal cancer screening.   The ambulatory clinic, established in 2016, has provided 204 individual patient encounters in its first 13 months, and has improved the patient experience by bridging barriers to care. SOSL reduces the cost of care to both the patients and the healthcare system by utilizing operating rooms outside of normal scheduled use, providing a completely volunteer workforce, and treating patients’ medical problems earlier and in the elective setting instead of allowing them to go untreated and eventually requiring hospital admission and more urgent and costly care.  Finally, this work also fosters an environment of overall wellness and resilience in over 500 volunteers from across departments and hospital institutions.  Of those volunteers surveyed, 58% report their SOSL work is more meaningful and fulfilling than their day-to-day work.

Conclusion: SOSL is a replicable model that achieves the Quadruple Aim by engaging the medical community in providing free surgical care to the underinsured.

69.02 The Overuse of Radioactive Iodine in Low-Risk Thyroid Cancer Patients

Posted on January 18, 2018

A. S. Moten1, H. Zhao2, A. I. Willis3  1Temple University Hospital,Department Of Surgery,Philadelphia, PA, USA 2Lewis Katz School Of Medicine At Temple University,Department Of Clinical Sciences, Section Of Biostatistics,Philadelphia, PA, USA 3Thomas Jefferson University,Department Of Surgery,Philadelphia, PA, USA

Introduction:  The 2015 American Thyroid Association (ATA) Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer suggests that radioactive iodine (RAI) ablation is not routinely needed for treatment of patients with low-risk thyroid cancer.  We sought to determine if disparities in the use of RAI ablation in low-risk thyroid cancer patients existed prior to publication of the ATA guidelines in order to identify patient groups who may be at risk of overtreatment with RAI after surgical treatment of low-risk thyroid cancer. 

Methods:  The Surveillance, Epidemiology and End Results (SEER) database was used to identify patients with papillary thyroid cancer diagnosed between 2011 and 2013.  Low-risk thyroid cancer was defined as T1, N0 or M0 disease.  Characteristics of patients with low-risk disease were compared to those who did not have low-risk disease.  Odds of presenting with low-risk disease were determined.  Among patients with low-risk disease, predictors of receiving RAI were determined.  Chi-square analyses were used to compare categorical variables, two-sample t tests to compare mean age at diagnosis, and logistic regression to determine odds ratios.  A p-value of ≤ 0.05 was used for all analyses to determine statistical significance.

Results: The study sample included 32,229 individuals, of which 17,286 (53.6%) had low-risk disease.  Mean (SD) age at diagnosis was 50.0 (14.8) years; 24,815 (77.0%) were female; and 21,318 (66.2%) were white.  Low-risk patients, compared to others, were older (mean age 51.3 versus 48.5 years; p-value < 0.001), more often female (81.6% versus 71.7%; p-value < 0.001), more often white (69.7% versus 62.0%; p-value < 0.001), and more often insured (87.6% versus 85.6%; p-value < 0.001).  Nearly 25% of low-risk patients were treated with RAI.  Predictors of overtreatment with RAI among low-risk patients included age less than 45 years (OR: 1.393; 95% CI: 1.250 – 1.552), age 45-64 years (OR: 1.275; 95% CI: 1.152 – 1.412), male sex (OR: 1.191; 95% CI: 1.086 – 1.305), Hispanic (OR: 1.236; 95% CI: 1.110 – 1.376) and Asian (OR: 1.306; 95% CI: 1.159 – 1.473) ethnicities, and more extensive lymph node removal surgery (OR: 1.243; 95% CI: 1.119 – 1.381).  Among low-risk patients, no patient who received RAI died of their disease, and only 5 (0.04%) patients who did not receive RAI died of their disease.  There was no survival benefit for patients who received RAI for low-risk disease compared to patients who did not receive RAI. 

Conclusion: Patients with low-risk thyroid cancer were more likely to receive RAI after surgery when not indicated under ATA guidelines if they were younger, male, Hispanic or Asian, or underwent extensive lymph node surgery.  Identification of groups at risk for overtreatment can help impact practice patterns and improve the effective utilization of healthcare resources.

69.03 Fertility Preservation Discussions in Premenopausal Breast Cancer Patients

Posted on January 18, 2018

A. J. Bartholomew1, M. L. Haslinger2, M. C. Masciello1, L. M. Bozzuto2, E. V. Tsiapali2  1Georgetown University Medical Center,School Of Medicine,Washington, DC, USA 2Georgetown University Medical Center,Department Of Surgery,Washington, DC, USA

Introduction:  Approximately 10% of new breast cancer diagnoses occur in premenopausal women annually. Oncologic management can involve systemic chemotherapy, endocrine therapy, or radiation depending on tumor characteristics. These treatments can compromise ovarian function, leading to premature menopause and impaired fertility. Thus, the importance of fertility preservation discussions (FPD) prior to treatment intervention is paramount. We aim to better characterize FPD in premenopausal breast cancer patients, to quantify utilization of fertility preservation care, and to identify patient demographics associated with low rates of FPD.

Methods:  A multi-institutional, retrospective chart review was performed on patients aged 18-40 with invasive breast cancer undergoing treatment from 2005 through 2015. Patient, tumor, and treatment characteristics were collected including age, ethnicity, parity, marital status, insurance, tumor staging, receptor status, and systemic treatment. Charts were reviewed for evidence of documented FPD, and, when identified, the specialty of the physician, timing of discussion in relation to treatment, and subsequent usage of fertility treatment were recorded.  Statistical analysis was performed to evaluate the influence of variables on whether a FPD occurred, and to investigate differences in counseling patterns between hospitals given a disparity in patient population demographics.

Results: Of the 227 premenopausal women in this cohort, only 103 (45.6%) had a documented FPD with their provider. Following discussions, 29 (28.2%) patients saw a fertility specialist prior to treatment. Of the FPD that occurred, 52% occurred prior to surgery, 96% before chemotherapy, and 95% prior to endocrine therapy. These discussions were initiated by the medical oncologist (82%), surgical oncologist (17%), and by an oncology nurse navigator (1%); 22% of patients had a FPD with multiple providers throughout treatment.  Of note, two medical oncologists accounted for 40% of the discussions. Multiparous women were significantly less likely to have a documented fertility discussion (p < 0.001). All other variables examined, including age, ethnicity, and insurance, were not significantly associated with FPD. Rates did not differ between hospitals, despite a significant baseline difference in ethnicity of the patient population.

Conclusion: An alarmingly low level of documented FPD persists in premenopausal women with invasive breast cancer. Women with more children were significantly less likely to receive fertility counseling from their providers.  A concerning 5% of discussions occurred after treatment that compromises fertility. These results highlight the need for improved provider attention to fertility preservation education in premenopausal breast cancer patients. Future interventions should consider the use of electronic medical record reminders and increased physician education to ensure that all patients are counseled prior to treatment.

 

 

68.10 Readmission Following Radical Cystectomy Is Common and Associated with Postdischarge Occurrences

Posted on January 18, 2018

I. Berger1, C. Wirtalla1, J. Ziemba2, T. Guzzo2, R. Kelz1  1University Of Pennsylvania,Department Of Surgery,Philadelphia, PA, USA 2University Of Pennsylvania,Department Of Urology,Philadelphia, PA, USA

Introduction: Efforts have been made to reduce post-operative readmissions due to their large burden on patients and the healthcare system. In spite of this, radical cystectomy remains a particularly morbid operation and with high rates of readmission. While patient demographics for radical cystectomy have failed to preoperatively identify high risk patients for readmission, postoperative occurrences are known to be a risk factor for readmission. However, their timing has not been appropriately described in radical cystectomy. In this study we sought to characterize postoperative occurrences with respect to their timing and association with readmission in order to target quality improvement interventions.

Methods: The National Surgical Quality Improvement Program database was queried for patients undergoing radical cystectomy from 2013-2015 using Common Procedural Terminology codes. Postoperative occurrences were identified and their impact on readmissions within 30 days of surgery was evaluated through multivariable analysis. Postoperative occurrences were characterized as predischarge or postdischarge, and their timing was described.

Results: 4,585 patients were identified with a mean length of stay of 8 days. Readmission rate was 24% and postoperative occurrence rate was 26%. There was a 62% rate of readmission in patients with any postoperative occurrence compared to 11% in patients that did not experience an occurrence (p<0.001). The greatest risk factor for readmission was the presence of a postdischarge occurrence (relative risk, 4.06 [95% CI, 3.97-4.15]). Readmission rates for patients with a predischarge versus a postdischarge occurrences were 28% and 79% (p<0.001). The median time of any postdischarge occurrence was 9 days. Urinary tract infection, sepsis, wound complication, and organ space infection were the most common postoperative occurrences with median times to occurrence of 10, 10, 8, and 10 days following discharge respectively. Frequency for each occurrence is shown in Figure 1.

Conclusion: Patients who experience a postoperative occurrence following radical cystectomy have a significantly higher risk of being readmitted. This risk is elevated even more if the event occurs after discharge. Infections are the most common form of postoperative occurrence and our study indicates that the majority occur before two weeks, a common time for a postoperative outpatient appointment. Outpatient interventions that target the first two weeks following discharge may be effective at identifying occurrences early and lowering readmission rates.

 

69.01 Lost in Translation: A Mixed Methods Pilot Study of Informed Consent in the Medical Mission Setting

Posted on January 18, 2018

L. A. Sceats1, R. Narayan1, A. Mezynski3, R. K. Woo2, A. M. Morris1, G. P. Yang1  1Stanford University,Surgery,Palo Alto, CA, USA 2University Of Hawaii,Pediatric Surgery,Honolulu, HI, USA 3Stanford University,S-SPIRE Center,Palo Alto, CA, USA

Introduction:
Interest in surgical missions to low- and middle-income countries (LMIC) is at an all-time high among surgeons in the United States. Ethical issues surround the care provided during these short-term trips, including whether informed consent is adequately conveyed. Little data exist about current practices, expectations, and limitations of informed consent on medical mission trips. In this pilot study, we examined physician communication of risk, translation adequacy, and patient comprehension among a population of adults undergoing inguinal hernia repair during a surgical mission trip in Guatemala.

Methods:
The study was conducted at a hospital in rural Guatemala that combines year round primary care with short-term surgical missions. All adult patients who were scheduled for inguinal herniorrhaphy in July 2017 agreed to participate. Within 10 minutes of a standard preoperative visit and consent conducted with a Spanish translator, patients underwent a separate structured interview covering sociodemographics and both closed- and open-ended questions to elicit comprehension. The informed consent process and post-visit interviews were audio-recorded, transcribed verbatim by a certified bilingual medical translator, independently coded by three researchers, and discussed to consensus. Using descriptive statistics and thematic analysis of open-ended responses, we integrated quantitative and qualitative methods focused on 1) clarity of information provided by surgeon, 2) accuracy of translation, and 3) patient understanding.

Results:
13 adult patients (mean age 49y, 76% male) were scheduled for inguinal herniorrhaphy; 61% had only a primary school education and 69% had no regular income. Surgeons uniformly conveyed standard risks of bleeding, infection, chronic pain and recurrence, but these were translated to patients only 49% of the time. All patients reported that they expected surgery would relieve pain and enable more activity. Within 10 minutes of the consent process, over 60% of patients indicated that they did not understand that postoperative recurrence or other complications were possible. Illustrative quotations are provided in Table 1.

Conclusion:
Despite visiting surgeons’ best attempts to provide adequate informed consent while on a foreign surgical mission, many patients did not comprehend the risks of surgery. Potential explanations include: low health literacy of patients, cultural disconnect with the concept of autonomy, surgeons’ delivery of information in large boluses of information, and failures of translation. Our pilot data reveal a critical need to develop more effective methods for communicating risk to patients about surgery completed as part of short-term medical missions to LMICs.
 

68.07 Do we CARE about the quality of case reports? A systematic assessment.

Posted on January 18, 2018

N. C. Dragnev1, S. L. Wong1  1Geisel School Of Medicine At Dartmouth,Surgery,Hanover, NH, USA

Introduction: Clinical case reports are important sources of information on the identification and treatment of new or rare diseases. Case reports are most useful when accurate, complete and transparent information are included. The CARE (CAse REport) Statement and Checklist represents consensus-based guidelines for clinical case reports. Despite adoption of CARE guidelines by the international EQUATOR Network, which represents medical journal editors, guideline developers, and other key stakeholders, how well case reports adhere to these guidelines is unknown.

Methods: A systematic PubMed and OVID search was used to identify case reports on isolated splenic metastasis from 2007-2017 in English language journals. MeSH search terms included “(isolated splenic metastasis OR solitary splenic metastasis) AND case report.” We retrieved 79 articles and 55 were found to directly address the topic of interest. Each eligible guideline was then scored using the 13 categories on the CARE checklist, which formally included 36 items ranging from the elements of presenting concerns to patient outcomes. Fulfillment of each item was rated on a dichotomous (yes/no) basis.

Results: Of the 55 case reports, none fully followed the CARE guidelines; only 56.4% of reports met 23 checklist items and none had more than 29 out of 36 total items (Figure). Only 35 of 55 (63.6%) case reports included the words “case report” in the title. In 22/55 (40.0%), patient symptoms were not described and in 26/55 (47.3%), the report’s abstract did not identify the main outcomes. All reports included the patient’s age and described the diagnostic work up. Most case reports reported the type of intervention (96.4%) and effect of the intervention (96.4%). Most also included a summary of the literature (49/55, 89.1%), but none included patient-assessed outcomes or the patient's perspective. In nearly all cases, a splenectomy was performed, followed by systemic chemotherapy. Only 27 reports (49.1%) included strengths and limitations of patient management, generally stating that the most effective treatment for isolated splenic metastasis is unknown because of its rare occurrence and lack of long-term evaluation of interventions. The rationale for this conclusion, including assessment of cause and effect, was not stated in 30.9% of case reports.

Conclusion: Based on a critical evaluation of case reports on isolated splenic metastasis, none completely followed the established CARE guidelines. Most reports did cover diagnostic workup and therapeutic interventions and gave a summary of the literature. Higher quality case reports would be useful in facilitating recognition of rare disease processes and informing clinical practice.

 

68.08 Missing Psychosocial Risk Factors is Associated with Increased Complications from Cancer Surgery

Posted on January 18, 2018

P. M. Meyers1, I. L. Leeds1, E. R. Haut1, J. E. Efron1, Z. O. Enumah1, F. M. Johnston1  1Johns Hopkins University School Of Medicine,Baltimore, MD, USA

Introduction:  Cancer surgery is known to carry additional risks compared to other surgical procedures and is often paired with comorbid therapies such as radiation and chemotherapy. Certain behavioral traits and inadequate social support may be risk factors for complications following cancer surgery. While psychosocial risk factors are often discussed in the context of cancer care, it is unclear if the medical record accurately captures them.

Methods:  In this study, adult surgical patients at a tertiary academic medical center were interviewed using validated risk screening instruments prior to curative abdominal cancer surgery. The questionnaire assessed 11 psychosocial risk factors including resilience, resourcefulness, substance abuse, and mental health. These same risk factors were assessed by retrospective chart review up to six months prior to surgery and compared to survey results. Concordance was calculated using positive and negative agreement and Cohen’s Kappa. Those positive by questionnaire but not the medical record were categorized as “clinically missed.” 30-day postoperative complications were abstracted from each patient’s medical record. Complication rates for those with and without missed psychosocial risk factors were statistically compared using Fisher’s exact test.

Results: We identified 57 eligible patients (61% male, median age of 65, IQR 55-72). Among those, 63% were identified as having 1 or more psychosocial risk factors by questionnaire while only 41% by chart review. Comparisons between chart review and patient survey revealed poor positive agreement among 9 of 11 risk factors (0-17%) and moderate agreement among 2 of 11 for depression (50%) and smoking status (75%). All showed excellent negative agreement (83-100%). Kappa analysis demonstrated poor-to-fair agreement among 10 of 11 risk factors (κ = 0 to 0.344).  Finalized 30-day outcomes were available for 40 patients with an overall complication rate of 37%. Seven of 11 psychosocial risk factors demonstrated a consistent nonsignificant trend for those with missed risk factors to have higher complication rates than those without (Figure). The complication rate for patients with at least one clinically missed psychosocial risk factor was 59% versus 17% in those without (p=0.015).

Conclusion: This study suggests a high level of discordance between formal screening and routine clinician interviewing and documentation for psychosocial risk factors. There is an association between missed risk factors and worse post-operative outcomes. Further work should demonstrate whether structured screening and documenting of psychosocial risk factors may better risk stratify patients preoperatively to allow for early intervention.
 

68.09 Disease-based Teams Improve Gastric Cancer Treatment Efficiency at a Large Safety Net Hospital

Posted on January 18, 2018

M. Ju1, S. C. Wang1, S. Syed2, D. Agrawal2, M. R. Porembka1  1University Of Texas Southwestern Medical Center,Surgery,Dallas, TEXAS, USA 2University Of Texas Southwestern Medical Center,Internal Medicine,Dallas, TEXAS, USA

Introduction:
The initiation of treatment for gastric cancer is a complex process requiring multiple diagnostic tests, staging evaluations, and invasive procedures. Inefficiencies in coordinating these tasks delay treatment initiation. Streamlining care through the development of disease-based teams (DBT) may reduce such delays through practice integration.

Methods:
A retrospective review of sequential gastric cancer patients treated at Parkland Hospital (Dallas, TX) between 2013 and 2015 was performed. Patients were divided into two groups: those who received care prior to the creation of a DBT and standardized care pathways (n=50), and those who received care after these protocols were implemented (n=50). Patients who underwent urgent resection or elected to pursue hospice care were excluded. Time to treatment (TTT) from initial endoscopy to initiation of treatment (chemotherapy or surgery) was measured and compared. The number of diagnostic tests performed, use of diagnostic laparoscopy, and treatment variability were also evaluated.

Results:
The two groups were demographically similar in terms of age, sex, stage distribution, tumor location, and setting of initial presentation (outpatient vs. emergency room). Implementation of a DBT was associated with a decrease in TTT from 84.1 +/-12.3 days to 32.5 +/-15.2 days (p < 0.02). This decrease was primarily related to parallel performance of subspecialty evaluations, staging imaging studies, and procedures. The number of staging tests decreased (3.8 vs. 2.2; p<0.05). The use of diagnostic laparoscopy in patients with clinically advanced local disease increased (18% vs. 94%, p<0.05).

Conclusion:
Creation of a gastric cancer-specific DBT and uniform care pathways at a large safety net hospital expedited the initiation of treatment, reduced unnecessary tests, and improved consistency in patient management.
 

68.06 Malignancies in Patients with Crohn's Disease: A US Population Based Study from NIS Database

Posted on January 18, 2018

S. Patil2, R. Chamberlain1  1Banner MD Anderson Cancer Center,Surgery,Gilbert, AZ, USA 2South Central Regional Medical Center,Surgery,Laurel, MS, USA

Introduction:  Cancer is a deadly sequelae of chronic inflammation in Crohn's disease (CD) patients, which follows the inflammation-dysplasia-carcinoma sequence. The inflammatory milieu increase the risk of colorectal cancer by 2-3 fold and small bowel cancer by up to 60 fold in CD patients compared to general population. The incidence of other cancer types in CD patients is poorly studied. This is the first population based study to estimate risk of GI and Extra-GI cancer in CD patients.

Methods:  National Inpatient Database (NIS 1998-2010) was used to identify patient with CD and demographic data including age, gender, race, mortality and type of malignancy were abstracted. Two groups were formed (CD patients with and without malignancies). The observed events for 13 randomly selected cancers were abstracted and converted to events/year/100,000 and compared against corresponding national incidence rate obtained from Surveillance Epidemiology and End Results (SEER) database. The Standardized Incidence Ratio (SIR) for each individual malignancy was calculated to assess the true risk. 

Results: 373,199 patients with CD were indentified in NIS database, of whom 29,971 (8.0%) suffered from at least one cancer. The mean age of CD patients with cancer was 62.3±16.0 years compared to those without cancer 46.8±19.0, p <0.001. More females compared to males developed CD (1.5: 1) and associated cancer (1.3: 1),   p <0.001. CD and associated cancer was highest in Caucasians compared to other ethnic groups, p <0.001. The mortality was higher in CD patients with cancer compared to those without cancer,  p <.001. Patients with CD are at higher risk of developing gallbladder cancer (SIR = 7364.3), anal canal cancer (SIR = 4772.6), extrahepatic biliary cancer (SIR = 4628.5), small bowel cancer (SIR = 3465.9), sarcoma (SIR = 2562.9), liver cancer (SIR = 979.9) , pancreas cancer (SIR = 654.7), rectal cancer (SIR = 649.2), colon cancer (SIR = 219.7), and lung cancer (SIR= 132.6), compared to general population. The risk for breast, stomach and esophageal cancer was lower in CD patients compared to general population.       

Conclusion: Patients with CD are at increased risk of developing GI and extra-GI cancers compared to general population, except for breast, stomach and esophageal cancers. The risk is higher in elderly, female and Caucasians. The mortality is higher in CD patients with cancer compared to those without cancer.      

 

68.04 Geography as a risk factor: the role of ZIP codes in predicting surgical oncology outcomes

Posted on January 18, 2018

A. N. Kothari1,2, S. A. Brownlee2, C. Fischer1, P. C. Kuo1,2, G. J. Abood1  1Loyola University Medical Center,Department Of Surgery,Maywood, IL, USA 2Loyola University Medical Center,One:MAP Division Of Clinical Informatics And Analytics,Maywood, IL, USA

Introduction:  Residential zip code can provide insight into patient socioeconomic status and community factors that influence perioperative care. Disparity in socioeconomic status has been well demonstrated to effect operative morbidity and mortality, but the role geography plays in outcomes of oncologic surgeries has yet to be established. The objective of this study was to determine the ability for residential zip code to predict postoperative outcomes in patients undergoing elective oncologic resections.

Methods: We conducted a retrospective cohort review using the Healthcare Cost and Utilization Project State Inpatient databases for Florida, Iowa, New York, and Washington. Included were patients that underwent open elective resection for the following malignancies: pancreas, colon, esophagus, stomach, and liver. The primary outcome was major inpatient morbidity or mortality. Regression-based predictive models were constructed using data from 2009 to 2012 (training set), with automated feature selection used to optimize fit. Zip code was added as a forced in variable when not selected for inclusion. Model performance was measured using in-sample data and validated using 2013 data. 

Results: 12,088 patients met inclusion criteria. Composite event rate: 19.9% (range across procedures: 18.7% – 21.1%). Event rate varied by residential zip code with a median event rate of 19.8 % (0.0 – 33.3%). Best fit regression model, without the inclusion of zip code, had an accuracy of 64.9% for predicting the primary outcome on the validation cohort. Zip code alone predicted the primary outcome in the validation cohort with an accuracy of 68.0%. Adding zip code to the best fit regression model increased the accuracy of prediction to 72.3%. Comparison of model performances is shown in Table.

Conclusion: Residential zip code can act as an independent predictor of postoperative outcomes. In addition, the inclusion of zip code can improve the performance of conventional models in the prediction of inpatient outcomes. Inclusion of residential zip code can offer an important adjunct for measuring risk-adjusted outcomes and identifying high risk geographic areas in order to optimize preoperative risk counseling.

68.05 Pilot Prehabilitation Program for Esophageal Cancer Patients During Neoadjuvant Therapy

Posted on January 18, 2018

L. C. Dewberry1, L. J. Wingrove3, A. Glode4, S. Jain2, M. Boniface1, S. L. Davis3, S. Leong3, K. Goodman2, S. Tracey2, W. T. Purcell3, M. D. McCarter1  1University Of Colorado Denver,Department of Surgery,Aurora, CO, USA 2University Of Colorado Denver,Department Of Radiation Oncology,Aurora, CO, USA 3University Of Colorado Denver,Division Of Medical Oncology,Aurora, CO, USA 4University Of Colorado Denver,Department Of Clinical Pharmacy,Aurora, CO, USA

Introduction:
Locally advanced esophageal cancer is a complex disease process often treated with neoadjuvant therapy followed by surgery. However, many patients experience a clinical deconditioning during neoadjuvant therapy. Prehabilitation programs in other areas of surgery have demonstrated improved postoperative outcomes. The aim of this study is to evaluate the feasibility of a pilot prehabilitation program and determine preliminary effects on surgical and cancer related outcomes.

Methods:
A retrospective review of patients treated at a single institution with resectable esophageal cancer was performed. Patients voluntarily undergoing the prehabilitation program (n=11) were compared to an equal number of historic controls (standard). Patients in the prehabilitation group received a protocol structured support in several clinical domains including nutrition, intravenous fluids, medications, psychosocial support, and physical exercise. Preliminary outcomes evaluated included nutritional parameters, surgical complications, mortality, readmission rates, and pathological outcomes.

Results:
Clinical stage and comorbidities were well matched between groups.  The structured prehabilitation program was feasible and well received by participants.  Fewer patients required admission during neoadjuvant therapy in the prehabilitation group (27.3% versus 54.5%). Percentage weight loss was 3% in the prehabilitation group versus 4.3% in the control group. Compared to the control group, the prehabilitation group demonstrated 0.0% versus 18.2% 30-day postoperative readmission rate and 18.2% vs. 27.3% 90-day postoperative readmission rate. The prehabilitation group had a similar percentage of complete pathologic response (27.0% versus 18.0%). There were no statistically significant differences between groups in regards to complications or mortality.

Conclusion:
The pilot prehabilitation program demonstrated its feasibility. Although the small population limits evaluation of statistical significance, trends in the data suggest a potential benefit of the prehabilitation program on neoadjuvant hospital admission rates, post surgical readmission rates, nutritional status, and oncologic outcomes. Larger randomized studies are warranted to assess the program’s overall utility.
 

68.03 Impact of Insurance Status on Receipt of Surgical Therapy and Outcomes in Early Stage Lung Cancer

Posted on January 18, 2018

S. M. Stokes1, E. Wakeam2, D. S. Swords1, J. R. Stringham3, T. K. Varghese3  1University Of Utah,Division Of General Surgery,Salt Lake City, UT, USA 2University Of Toronto,Division Of Thoracic Surgery,Toronto, ON, Canada 3University Of Utah,Division Of Cardiothoracic Surgery,Salt Lake City, UT, USA

Introduction: Insurance coverage by government funded programs has been a topic of debate, but the impact on outcomes in the modern era of evidence-based guidelines and increasing specialization remains unclear. Previous work have suggested that outcomes for Medicaid patients are inferior to those of privately insured patients. We sought to examine differences in receipt of cancer therapy and outcomes for early stage non-small lung cancer (NSCLC) patients according to their insurance coverage.

Methods:  Clinical, T1-3, N0-1 NSCLC cases were identified in the 2004-2014 National Cancer Database. Patients were compared across four insurance groups: Private, Medicare, Medicaid, and Uninsured. Patients with unknown or other insurance status were excluded. A multivariable, linear regression model was used to examine the effects of insurance status on time to definitive surgical therapy by adjusting for patient and facility characteristics. Receipt of surgery, radiation, and chemotherapy were examined with multivariable logistic regression. Survival analysis was conducted with Cox regression to delineate the impact of insurance status on post-treatment survival.

Results: There were 291,732 patients presenting with early NSCLC (76,908 Private, 196,740 Medicare, 12,896 Medicaid, and 5,188 Uninsured). After adjusting for patient and facility characteristics, Medicaid and uninsured patients received definitive surgical therapy significantly later than privately insured patients (9.4 days and 7.5 days respectively, p < 0.001) and were significantly less likely to receive surgery (OR 0.50, 95% CI 0.48-0.53 and OR 0.48, 95% CI 0.45-0.52). Among all patients, 6.14% did not receive any form of treatment. Uninsured patients were more likely to receive no treatment (OR 2.26, 95% CI 2.02-2.53), followed by Medicaid patients (OR 1.75, 95% CI 1.61-1.90). Thirty and 90-day mortality were worse in Medicare, Medicaid, and uninsured populations. Overall survival was significantly worse in the Medicaid and uninsured populations (HR 1.44, 95% CI 1.39-1.48 and HR 1.37, 95% CI 1.31-1.43). 

Conclusion: Even in the modern era, uninsured and Medicaid patients with early stage NSCLC have decreased odds of definitive treatment and poor outcomes after treatment compared to privately insured patients. This may relate to inferior access and underuse of cancer therapy in this population. Given significant state and federal expenditures on the Medicaid program, strategies for improving the optimal treatment and outcomes of Medicaid patients with lung cancer are needed.

 

68.01 Trends in Inpatient Palliative Care Referrals by Type of Malignancy

Posted on January 18, 2018

J. M. Ruck1, J. K. Canner1, T. J. Smith2, F. M. Johnston1  1Johns Hopkins School Of Medicine,Department Of Surgery,Baltimore, MD, USA 2Johns Hopkins Hospital,Sidney Kimmel Comprehensive Cancer Center,Baltimore, MD, USA

Introduction:
The field of oncology has led palliative care (PC) utilization and demonstrated how PC improves patient quality of life. Yet, the frequency of and factors associated with PC use during oncology-related hospitalizations remain unknown.

Methods:
Using the National Inpatient Sample dataset, we identified 124,186 hospitalizations 2012-2014 for patients with a high risk of in-hospital mortality (DRG risk of mortality=3-4) and a primary diagnosis of malignancy (melanoma, breast, colon, gynecologic, prostate, male genitourinary (GU), head/neck, urinary tract, non-colon gastrointestinal, lung, brain, bone/soft tissue, endocrine, or non-lung thoracic). Univariate analyses were stratified by malignancy type. PC use was identified using the V66.7 ICD-9 code. Change in PC use over time was assessed using linear regression. The number of secondary diagnoses, median cost of hospitalization, and frequency of palliative procedures, surgical procedures, and in-hospital death were compared by PC use. Patient factors associated with the cost of hospitalization were identified using multivariable linear regression.

Results:
PC use increased 2012-2014 for all malignancy types except brain cancer. Patients utilizing PC had more secondary diagnoses than those who did not (median 12-17 vs. 11-15, all p<0.001), a higher frequency of palliative procedures (4-36% vs. 0-35%, all p<0.01 except non-lung thoracic), a lower frequency of operative procedures (4-33% vs. 34-79%, all p<0.001), a higher rate of in-hospital death (30-45% vs. 4-10%, all p<0.001), a lower total cost for the hospitalization (median, in thousands: $14-61 vs. $33-98, all p<0.01 except male GU). PC was associated with a shorter or similar – but not longer – length of stay vs. no PC. In an adjusted analysis, the cost of hospitalization was associated with patient sex (female vs. male, $5,248 lower), race (African American vs. other, $12,773 lower), age (per year older, $720 lower), operative procedure(s) (had vs. didn’t have, $33,149 higher), in-hospital death (died vs. alive, $26,269 higher), length of stay (per day longer, $9,965 higher), and PC (PC vs. no PC, $13,191 lower) (all p<0.001).

Conclusion: 
In summary, inpatient PC utilization has increased for patients with a high predicted risk of in-hospital mortality, though PC was disproportionately used for patients who experience in-hospital death. PC was associated with lower utilization of surgical procedures, shorter length of stay, and lower hospitalization cost. Lower hospitalization cost is also seen for patients who are older, female, or African American, suggesting possible disparities in cancer care utilization by age, sex, and race.
 

68.02 Racial Disparity in Preoperative Chemotherapy Use in Gastric Cancer Patients in the United States

Posted on January 18, 2018

N. Ikoma1,2, J. Cormier1, B. Feig1, X. L. Du2, J. Yamal2, P. Das1, J. A. Ajani1, C. Roland1, K. Fournier1, R. Royal1, P. Mansfield1, B. Badgwell1  1University Of Texas MD Anderson Cancer Center,Houston, TEXAS, USA 2University Of Texas Health Science Center At Houston,Houston, TEXAS, USA

Introduction:
Racial disparity is widely reported in gastric cancer in the United States. The use of preoperative chemotherapy in patients with resectable gastric cancer has sharply increased over the past 10 years; however, no studies have investigated whether race/ethnicity is associated with the use of preoperative chemotherapy or subsequent outcomes in gastric cancer.

Methods:
Patients with clinical T2-4bN0-1M0 gastric adenocarcinoma, as defined by the AJCC 8th edition, who underwent gastrectomy during 2006-2014 were identified from the National Cancer Database. The main exposure variable was race/ethnicity, and the main outcome variable was preoperative chemotherapy use, defined by systemic therapy surgery sequence code. Multiple logistic regression was conducted to examine factors associated with preoperative chemotherapy use. A multiple Cox regression model was used to examine overall survival, as a secondary analysis.

Results:
We identified 16,945 patients who met study criteria, of whom 8,286 (49%) underwent preoperative chemotherapy. The median age was 65 years (interquartile range 57-74 years), and 69% were male; 71% were non-Hispanic (NH) white, 12% were NH black, 9% were Hispanic, and 7% were Asian and Pacific Islander (API). The use of preoperative chemotherapy remarkably increased over the study period, from 34% in 2006 to 65% in 2014. Preoperative chemotherapy was more commonly used in cardia tumors than in non-cardia tumors (83% vs. 44%, in 2014). In multivariable analysis, NH blacks (odds ratio [OR] 0.76, 95% confidence interval [CI] 0.67-0.86; p<0.001), Hispanics (OR 0.83, 95% CI 0.72-0.95; p=0.006), and APIs (OR 0.62, 95% CI 0.53-0.72; p<0.001) were associated with less frequent use of preoperative chemotherapy compared with NH whites. Increases in preoperative chemotherapy use over time were homogeneous between race/ethnicity groups. Insurance status, education level, and treatment at non-academic hospital mediated an enhanced effect of racial/ethnic disparity in preoperative chemotherapy use. In Cox regression models, use of preoperative chemotherapy and radiation therapy was associated with reduced racial/ethnic disparity in overall survival.

Conclusion:
Racial/ethnic disparity in the use of preoperative chemotherapy and outcomes exists among gastric cancer patients in the United States.  Race was independently associated with low frequency of preoperative chemotherapy use, and no insurance status, low education level, and treatment at non-academic hospital mediated an enhanced effect of the disparity in preoperative chemotherapy. Efforts to improve the access to high-quality cancer care in minority groups may reduce racial disparity in gastric cancer in the United States.
 

« Previous Page
Next Page »