Testing the feasibility of mobile ecological momentary assessment for symptom burden and management among metastatic cancer patients.

Background: Individuals who have metastatic cancer experience substantial physical and psychological distress (e.g., pain, depression, anxiety) from their disease and its treatment compared to patients with less advanced disease. As the burden of symptoms varies over time, ecological momentary assessment (EMA) may be used to better understand patients' symptom trajectories, complimenting traditional longitudinal data collection methods. However, few have used EMA in patients with metastatic disease. The current study adds to the existing literature by exploring interrelated, common cancer-related symptoms of pain, anxiety, and depression and use of cannabis-based products, opioid medications, other (nonopioid) pain medications, and medications for anxiety or depression. Methods: An eight-day prospective observational feasibility study was conducted among 50 patients with metastatic cancer recruited from seven solid cancer clinics at The Ohio State University Comprehensive Cancer Center. Participants completed a week of interval-contingent mobile EMA, administered daily at 9 a.m., 3 p.m., and 8 p.m., and a comprehensive interviewer-administered questionnaire on Day 8. Participants were queried on their symptom burden and management strategies (i.e., use of medications and cannabis). We considered EMA to be feasible if a priori retention (80%) and adherence goals (75%) were met. Results: Seventy-nine percent of eligible patients contacted enrolled in the study (n = 50 of 63). Among those enrolled, 92% were retained through Day 8 and 80% completed >90% of EMAs, exceeding a priori objectives. Participants' average pain, anxiety, and depressive symptoms across the week of EMA ranged from 1.7 to 1.8 (1 to 5 scale). Symptoms varied little by day or time of administration. On Day 8, significant proportions of participants reported past-week use of medications and cannabis for symptom management. Conclusions: Participants exceeded a priori adherence and retention objectives, indicating that mobile EMA is feasible among metastatic cancer patients, addressing a gap in the existing literature and informing future research. Restricting eligibility to participants with a minimum cutoff of symptom burden may be warranted to increase observations of symptom variability and provide opportunities for future health interventions. Future research is needed to test the acceptability and quality of data over a longer study period in this patient population.

Alpelisib-Induced Diabetic Ketoacidosis: A Pharmacovigilance Analysis of the FDA Adverse Event Reporting System and Review of the Literature.

BACKGROUND: Alpelisib is a PI3K inhibitor indicated with fulvestrant for treatment of advanced or metastatic hormone receptor (HR)-positive, human epidermal growth factor receptor 2 (HER2)-negative, PIK3CA-mutated breast cancer. In the phase III SOLAR-1 trial, grade 3/4 hyperglycemic events were reported in 36.6% of patients receiving alpelisib-fulvestrant compared to 0.7% receiving placebo-fulvestrant. As case reports of diabetic ketoacidosis (DKA) have been associated with alpelisib use, the goal of this study was to characterize the FAERS reported cases of this severe adverse effect. METHODS: A retrospective disproportionality analysis was performed using the FAERS database by calculating the reporting odds ratio (ROR) of DKA events with alpelisib from 2019 to 2022. A PubMed literature review of case reports characterizing alpelisib-induced DKA was performed. RESULTS: Pharmacovigilance database analysis revealed significance in reporting among 87 DKA cases with alpelisib (ROR 9.84, 95% confidence interval 7.3-13.2), including hospitalization and death as reported outcomes. Review of 11 published case reports reveals median onset of DKA at 14 days with successful rechallenge possible. CONCLUSION: Significant association with reporting exists between DKA and alpelisib exposure. We observed similar median time to onset of hyperglycemia between our analysis compared to that reported in SOLAR-1. Considering early onset of this toxicity, it is imperative that patients be closely monitored when initiating alpelisib. Addition of a preemptive antihyperglycemic or escalation in those previously on antihyperglycemic medications is beneficial in decreasing the severity of hyperglycemia with alpelisib. Further study investigating risk factors is warranted to better elucidate which patients require preemptive therapy.

Current cannabis use and pain management among US cancer patients.

BACKGROUND: National studies reporting the prevalence of cannabis use have focused on individuals with a history of cancer without distinction by their treatment status, which can impact symptom burden. While pain is a primary motivation to use cannabis in cancer, the magnitude of its association with cannabis use remains understudied. METHODS: We examined cannabis use and pain management among 5523 respondents of the Behavioral Risk Factor Surveillance System with a cancer history. Survey-weighted prevalence proportions of respondents' cannabis use are reported, stratified on cancer treatment status. Regression models estimated odds ratios (ORs) and 95% confidence intervals (CIs) of cancer-related pain and cannabis use. RESULTS: Cannabis use was slightly more prevalent in those undergoing active treatment relative to those who were not undergoing active treatment (9.3% vs. 6.2%; P=0.05). Those under active treatment were more likely to use cannabis medicinally (71.6% vs. 50.0%; P=0.03). Relative to those without cancer-related pain, persons with pain under medical control (OR 2.1, 95% CI, 1.4-3.2) or uncontrolled pain were twice as likely to use cannabis (OR 2.0, 95% CI, 1.1-3.5). CONCLUSIONS: Use of cannabis among cancer patients may be related to their treatment and is positively associated with cancer-related pain. Future research should investigate the associations of cannabis use, symptom burden, and treatment regimens across the treatment spectrum to facilitate interventions.

Associations of Cancer Patients' Demographic and Clinical Characteristics With Cannabis-Related Interest and Behaviors.

Background: Cannabis interest and use is increasing in the United States, yet research on its use among cancer patients is limited. Methods: Individuals with cancer completed an anonymous cross-sectional questionnaire. Multivariable logistic regressions estimated odds ratios (OR) between patients' demographic and clinical characteristics with cannabis-related interest, current use, and provider recommendation. Results: Participants (n = 943) were, on average, 61.7 years old. Older patients were less likely to use cannabis products (OR = 0.42, confidence interval [95% CI]: 0.26-0.69) and less likely to be interested in cannabis (OR = 0.60, 95% CI: 0.44-0.84) than younger patients. Those with higher education were less likely to be using cannabis (OR = 0.41, 95% CI: 0.25-0.67) and less likely to have received a provider recommendation of cannabis use than the least educated (OR = 0.38, 95% CI: 0.19-0.76). Cancer spread and type were significant correlates of provider recommendation of cannabis use. Conclusions: Additional research is warranted to better understand cancer patients' motivations for cannabis use and interest.

Marijuana and Cannabidiol Use Prevalence and Symptom Management Among Patients with Cancer.

Symptoms such as pain, nausea, and anxiety are common in individuals with cancer. Treatment of these issues is often challenging. Cannabis products may be helpful in reducing the severity of these symptoms. While some studies include data on the prevalence of cannabis use among patients with cancer, detailed data remain limited, and none have reported the prevalence of cannabidiol (CBD) use in this population. Adult patients with cancer attending eight clinics at a large, NCI-designated Comprehensive Cancer Center completed a detailed, cannabis-focused questionnaire between 2021 and 2022. Eligible participants were diagnosed with invasive cancer and treated in the past 12 months. Summary statistics were calculated to describe the sample regarding cannabis use. Approximately 15% (n = 142) of consented patients (n = 934) reported current cannabis use (defined as use within the past 12 months). Among which, 75% reported cannabis use in the past week. Among current cannabis users, 39% (n = 56; 6% overall) used CBD products. Current users reported using cannabis a median of 4.5 (interquartile range: 0.6­7.0) days/week, 2.0 (1.0­3.0) times per use/day, and for 3 years (0.8­30.0). Use patterns varied by route of administration. Patients reported moderate to high relief of symptoms with cannabis use. This study is the most detailed to date in terms of cannabis measurement and provides information about the current state of cannabis use in active cancer. Future studies should include complete assessments of cannabis product use, multiple recruitment sites, and diverse patient populations. SIGNIFICANCE: Clinicians should be aware that patients are using cannabis products and perceive symptom relief with its use.

Use of patient-reported outcome measures for oncology drugs receiving accelerated approval.

Patient-reported outcomes (PROs) represent an important evaluation of health-related quality of life that has become more commonly incorporated into oncology drug clinical trials. The frequency of PRO inclusion as an endpoint in oncology drug clinical trials leading to the initial accelerated approval of a new therapy is not yet known. We conducted a cross-sectional study evaluating all new drug applications submitted to the FDA over the past 10 years (2013-2022) that led to the initial approval of an oncology drug through the accelerated approval process. The objective was to assess whether the trials leading to such an approval included PROs. Between 2013 and 2022, the FDA approved 59 unique drugs for an oncology indication via the accelerated approval pathway, and 35 (59%) included a PRO assessment in the clinical trial. A median of 1 PRO measurement was used in each trial, with 23 different types of PRO assessment tools were used across the 59 new drug applications. In summary, we found that PRO measurements are inconsistently utilized in trials leading to initial accelerated approval of oncology drugs, and there seems to be a lack of harmonization of different PRO measurement tools used across trials.

Hepatitis B reactivation in patients with multiple myeloma treated with anti-CD38 monoclonal antibody-based therapies: a pharmacovigilance analysis.

BACKGROUND: Daratumumab and isatuximab are anti-CD38 monoclonal antibodies indicated for the treatment of multiple myeloma. These agents can increase the risk of infectious complications, including viral infections. Cases of hepatitis B virus (HBV) reactivation have been reported in the literature in patients receiving anti-CD38 monoclonal antibody-based therapies. AIM: The objective of this analysis was to determine if the association between anti-CD38 monoclonal antibody exposure and the development of hepatitis B reactivation had a detectable reporting signal in the United States Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS). METHOD: We conducted a post marketing pharmacovigilance analysis by querying the FAERS for reports of HBV reactivation with daratumumab or isatuximab exposure reported between 2015 and 2022. Disproportionality signal analysis was conducted by calculating reporting odds ratios (RORs). RESULTS: Sixteen cases of hepatitis B virus reactivation were reported in the FAERS database among patients receiving daratumumab or isatuximab reported between 2015 and 2022. The ROR for HBV reactivation was statistically significant for both daratumumab (ROR 4.76, 95% CI 2.76-8.22) and isatuximab (ROR 9.31, 95% CI 3.00-28.92). CONCLUSION: Overall, our analysis demonstrates a significant reporting signal for HBV reactivation with daratumumab and isatuximab.

Acquired thrombotic thrombocytopenic purpura associated with immune checkpoint inhibitors: A real-world study of the FDA adverse event reporting system.

INTRODUCTION: Immune checkpoint inhibitors (ICIs) are used for a variety of cancers and are associated with a risk of developing immune-related adverse events, most commonly colitis, dermatitis, hepatitis, and thyroiditis. Rare autoimmune hematologic toxicities have been reported but are less well-described in the literature. Acquired thrombotic thrombocytopenic purpura (TTP) is a life-threatening autoimmune condition that has been reported with ICIs but has been limited to case reports. METHODS: We performed a retrospective observational analysis of the United States Food and Drug Administration Adverse Event Reporting System (FAERS) data. We searched for cases of TTP reported with exposure to ICIs from initial FDA approval for each agent to December 31, 2021. Disproportionality signal analysis was done by calculating the reporting odds ratio (ROR). RESULTS: There were 35 reports of TTP with ICIs in the FAERS database, including atezolizumab (n = 7), durvalumab (n = 2), nivolumab (n = 18), and pembrolizumab (n = 8). The ROR was significant for atezolizumab (ROR 6.22, 95% CI 2.96-13.09), nivolumab (ROR 3.16, 95% CI 1.99-5.03), and pembrolizumab (ROR 2.56, 95% CI 1.28-5.12). CONCLUSIONS: There is a significant reporting signal of TTP with several ICI agents. Clinicians should be aware of and monitor for signs of this potentially serious adverse event.

Opioid Taper Practices Among Clinicians.

INTRODUCTION: Opioid dose tapers are used frequently when cross-titrating from one or more opioids to another or when discontinuing therapy. Currently, there is no universally accepted evidence-based standard of care for this procedure which can leave patients at risk for withdrawal symptoms, inadequate pain control, or elevated suicide risk. OBJECTIVE: The objective of this study was to examine practices and rationale among clinicians, to determine if there is a difference among respondents in their comfort level, method and rationale for tapering opioids at various morphine milligram equivalents (MME) and to assess the need for the development of a standard of care. METHODS: Data were derived from an electronic survey developed using SurveyMonkey®. The survey was disseminated via e-mail listservs, social media, and professional organizations. Data were collected regarding profession, confidence tapering opioids at varying total MME, method and rationale for tapering, and pharmacologic management of withdrawal symptoms. Pearson's Chi squared and Fisher's exact tests were used to assess statistical significance of results. RESULTS: A total of 149 clinicians completed the survey, physicians, NPs, pharmacists, and PAs accounted for 51%, 20%, 19%, and 10% of participants, respectively. Overall, 55% of the respondents self-identified as pain specialists. There were no statistically significant differences in reported comfort level among the different types of providers. Nearly 50% of participants indicated their rationale for tapering or discontinuing opioids was the 2016 CDC guidelines. CONCLUSION: Despite that the majority of providers surveyed self-identified as pain specialists, over 50% were not comfortable tapering opioids at doses greater than 120 MME/day. This observation suggests a need for further education and establishment of consensus guidelines on method and rationale for opioid tapering. Provider motivation for tapering was largely influenced by CDC guidelines based on low quality evidence. This strengthens the argument for the creation of guidelines based on high quality evidence.

In-house intensivist presence does not affect mortality in select emergency general surgery patients.

BACKGROUND: This study aimed to assess the relationship between availability of round-the-clock (RTC) in-house intensivists and patient outcomes in people who underwent surgery for a life-threatening emergency general surgery (LT-EGS) disease such as necrotizing soft-tissue infection, ischemic enteritis, perforated viscus, and toxic colitis. METHODS: Data on hospital-level critical care structures and processes from a 2015 survey of 2,811 US hospitals were linked to patient-level data from 17 State Inpatient Databases. Patients who were admitted with a primary diagnosis code for an LT-EGS disease of interest and underwent surgery on date of admission were included in analyses. RESULTS: We identified 3,620 unique LT-EGS admissions at 368 hospitals. At 66% (n = 243) of hospitals, 83.5% (n = 3,021) of patients were treated at hospitals with RTC intensivist-led care. These facilities were more likely to have in-house respiratory therapists and protocols to ensure availability of blood products or adherence to Surviving Sepsis Guidelines. When accounting for other key factors including overnight surgeon availability, perioperative staffing, and annual emergency general surgery case volume, not having a protocol to ensure adherence to Surviving Sepsis Guidelines (adjusted odds ratio, 2.10; 95% confidence interval, 1.12-3.94) was associated with increased odds of mortality. CONCLUSION: Our results suggest that focused treatment of sepsis along with surgical source control, rather than RTC intensivist presence, is key feature of optimizing EGS patient outcomes. LEVEL OF EVIDENCE: Therapeutic, level III.

Hospital Admissions Associated With Cancer Pain in Older Adults With and Without Dementia.

BACKGROUND: Neoplasm-related pain is often suboptimally treated, contributing to avoidable suffering and increased medical resource use and costs. We hypothesized that dementia may contribute to increased resource use and costs in patients hospitalized for neoplasm-related pain in the United States. AIMS: To examine how persons with cancer and dementia use medical resources and expenditures in US hospitals compared to ondividuals without dementia. DESIGN: This study examined a retrospective cohort. SETTING: Admissions to US hospitals for neoplasm-related pain from 2012-2016 PARTICIPANTS/SUBJECTS: METHODS: Data were obtained from the 2012-2016 National Inpatient Sample (NIS). The sample included hospital admissions of individuals aged 60 or older with a primary diagnosis of neoplasm-related pain. Dementia was defined using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), and ICD-10-CM diagnosis codes. Primary outcomes were number of admissions, costs, and length of stay (LOS). Descriptive statistics and multivariable regression models were used to examine the relationships among dementia, costs, and LOS. RESULTS: Of 12,034 admissions for neoplasm-related pain, 136 (1.1%) included a diagnosis of dementia and 11,898 (98.9%) did not. Constipation was present in 13.2% and 24.5% of dementia and nondementia admissions, respectively. The median LOS was 4 days in persons with dementia and three in those without. Mean costs per admission were higher in persons without dementia ($10,736 vs. $9,022, p = .0304). In adjusted regression results, increased costs were associated with nonelective admissions and longer LOS, and decreased costs with age above the mean. In contrast, decreased LOS was associated with age above the mean and nonelective admissions. Dementia was associated with neither endpoint. CONCLUSION: This study provides nurses and other health care professionals with data to further explore opportunities for improvement in cancer pain management in patients with and without dementia that may optimize use of medical resources.

Opioid use by cancer status and time since diagnosis among older adults enrolled in the Prostate, Lung, Colorectal, and Ovarian screening trial in the United States.

BACKGROUND: Dosing limits in opioid clinical practice guidelines in the United States are likely misapplied to cancer patients, however, opioid use may be difficult to ascertain as they are largely excluded from opioid use studies. METHODS: The primary objective was to determine whether cancer patients were more likely to be chronic opioid users after diagnosis. We described prescription opioid use among U.S. older adult cancer patients during two time periods, within 2 years of diagnosis (short-term) and at least 2 years beyond diagnosis (long-term), compared to those without cancer (controls). Among participants in the Prostate, Lung, Colorectal, and Ovarian (PLCO) screening trial with linkages to Medicare Part D data during 2011-2015, we used multivariable logistic regression to estimate the association between cancer diagnosis and opioid use outcomes controlling for demographics. The primary outcome of opioid use was measured with the following metrics: Any opioid use, chronic use (90 consecutive days supply of opioid use while allowing for a 7-day gap between refills), high use (average daily morphine equivalent (MME) ≥120 mg for any 90-day period), and total MME dose above 2,000 mg (MME2000 ). RESULTS: The short-term cohort included 1,491 cancer patients and 24,930 controls. Any use in the 2-year post-diagnosis period was higher among cancer patients OR 3.3 (95% CI: 3.0-3.7). Chronic use rates were similar by cancer status (4.6% vs. 3.8% for cases and controls, respectively). The long-term cohort included 4,377 cancer patients and 27,545 controls. Rates of any use were similar among cancer patients and controls (63% vs. 59%). CONCLUSIONS: Any opioid use was similar among long-term cancer survivors compared to controls, but differed among short-term survivors for any opioid use and marginally for chronic opioid use.

Importance of Radiologists in Optimizing Outcomes for Older Americans with Acute Abdomen.

BACKGROUND: Patients presenting with acute abdominal pain often undergo a computed tomography (CT) scan as part of their diagnostic workup. We investigated the relationship between availability, timeliness, and interpretation of CT imaging and outcomes for life-threatening intra-abdominal diseases or "acute abdomen," in older Americans. METHODS: Data from a 2015 national survey of 2811 hospitals regarding emergency general surgery structures and processes (60.1% overall response, n = 1690) were linked to 2015 Medicare inpatient claims data. We identified beneficiaries aged ≥65 admitted emergently with a confirmatory acute abdomen diagnosis code and operative intervention on the same calendar date. Multivariable regression models adjusted for significant covariates determined odds of complications and mortality based on CT resources. RESULTS: We identified 9125 patients with acute abdomen treated at 1253 hospitals, of which 78% had ≥64-slice CT scanners and 85% had 24/7 CT technicians. Overnight CT reads were provided by in-house radiologists at 14% of hospitals and by teleradiologists at 66%. Patients were predominantly 65-74 y old (43%), white (88%), females (60%), and with ≥3 comorbidities (67%) and 8.6% died. STAT radiology reads by a board-certified radiologist rarely/never available in 2 h was associated with increased odds of systemic complication and mortality (adjusted odds ratio 2.6 [1.3-5.4] and 2.3 [1.1-4.8], respectively). CONCLUSIONS: Delays obtaining results are associated with adverse outcomes in older patients with acute abdomen. This may be due to delays in surgical consultation and time to source control while waiting for imaging results. Processes to ensure timely interpretation of CT scans in patients with abdominal pain may improve outcomes in high-risk patients.

The Need for Accurate Risk Assessment in a High-Risk Patient Population: A NSQIP Study Evaluating Outcomes of Cholecystectomy in the Patient With Cancer.

BACKGROUND: Cholecystectomy is considered a low-risk procedure with proven safety in many high-risk patient populations. However, the risk of cholecystectomy in patients with active cancer has not been established. METHODS: The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database was queried to identify all patients with disseminated cancer who underwent cholecystectomy from 2005 to 2016. Postcholecystectomy outcomes were defined for patients with cancer and those without by comparing several outcomes measures. A multivariate model was used to estimate the odds of 30-d mortality. RESULTS: We compared outcomes in 3097 patients with disseminated cancer to a matched cohort of patients without cancer. Patients with cancer had more comorbidities at baseline: dyspnea (10.5% versus 7.0%, P < 0.0001), steroid use (10.1% versus 3.0%, P < 0.0001), and loss of >10% body weight in 6-mo prior (9.3% versus 1.6%, P < 0.0001). Patients with cancer sustained higher rates of wound (2.3% versus 5.6%, P < 0.0001), respiratory (1.4% versus 3.9%, P < 0.0001), and cardiovascular (2.0% versus 6.8%, P < 0.0001) complications. In addition, patients with disseminated cancer experienced a longer length of stay and higher 30-d mortality. Multivariate modeling showed that the odds of 30-d mortality was 3.3 times greater in patients with cancer. CONCLUSIONS: Compared to patients without cancer, those with disseminated cancer are at higher risk of complication and mortality following cholecystectomy. Traditional treatment algorithms should be used with caution and care decisions individualized based on the patient's disease status and treatment goals.

Re-examining "Never Letting the Sun Rise or Set on a Bowel Obstruction" in the Era of Acute Care Surgery.

BACKGROUND: Small bowel obstruction (SBO) no longer mandates urgent surgical evaluation raising the question of the role of operating room (OR) access on SBO outcomes. METHODS: Data from our 2015 survey on emergency general surgery (EGS) practices, including queries on OR availability and surgical staffing, were anonymously linked to adult SBO patient data from 17 Statewide Inpatient Databases (SIDs). Univariate and multivariable associations between OR access and timing of operation, complications, length of stay (LOS), and in-hospital mortality were measured. RESULTS: Of 32,422 SBO patients, 83% were treated non-operatively. Operative patients were older (median 66 vs 65 years), had more comorbidities (53% vs 46% with ≥ 3), and experienced more systemic complications (36% vs 23%), higher mortality (2.8% vs 1.4%), and longer LOS (median 10 vs 4 days). Patients had lower odds of operation if treated at hospitals lacking processes to tier urgent cases (aOR 0.90, 95% CI [0.83-0.99]) and defer elective cases (aOR 0.87 [0.80-0.94]). Patients had higher odds of operation if treated at hospitals with surgeons sometimes (aOR 1.14 [1.04-1.26]) or rarely/never (aOR 1.16 [1.06-1.26]) covering EGS at more than one location compared to always. Odds of systemic complication (OR 2.0 [1.6-2.4]), operative complication (OR 1.5 [1.2-1.8]), and mortality were increased for very late versus early operation (OR 2.6 [1.7-4.0]). CONCLUSIONS: Although few patients with SBO require emergency surgery, we identified EGS structures and processes that are important for providing timely and appropriate intervention for patients whose SBO remains unresolved and requires surgery.

Leaving Against Medical Advice After Emergency General Surgery: Avoiding a Two-Hit Effect.

BACKGROUND: Emergency general surgery has higher adverse outcomes than elective surgery. Patients leaving the hospital against medical advice (AMA) have a greater risk for readmission and complications. We sought to identify clinical and demographic characteristics along with hospital factors associated with leaving AMA after EGS operations. METHODS: A retrospective review of the Nationwide Inpatient Sample was performed. All patients who underwent an EGS procedure accounting for >80% of the burden of EGS-related inpatient resources were identified. 4:1 propensity score analysis was conducted. Regression analyses determined predictive factors for leaving AMA. RESULTS: 546,856 patients were identified. 1085 (0.2%) patients who underwent EGS left AMA. They were more likely to be men (59% versus 42%), younger (median age 51 y, IQR [37.61] versus 54, IQR [38.69]), qualify for Medicaid (26% versus 13%) or be self-pay (17% versus 9%), and be within the lowest quartile median household income (40% versus 28%) (all P < 0.05). After applying 4:1 propensity score matching, individuals who were self-pay (OR 3.15, 95% CI 2.44-4.06) or insured through Medicare (OR 2.75, 95% CI 2.11-3.57) and Medicaid (OR 3.58, 95% CI 2.83-4.52) had increased odds of leaving AMA compared with privately insured patients. In addition, history of alcohol (OR 2.21, 95% CI 1.65-2.98), drug abuse (OR 4.54, 95% CI 3.23-6.38), and psychosis (OR 2.31, 95% CI 1.65-3.23) were associated with higher likelihood for leaving AMA. CONCLUSIONS: Patients undergoing EGS have a high risk of complications, and leaving AMA further increases this risk. Interventions to encourage safe discharge encompassing surgical, psychiatric, and socioeconomic factors are warranted to prevent a two-hit effect and compound postoperative risk.