Safety Net Health Care System Research Articles

A cross-sectional analysis of factors associated with stress, burnout and turnover intention among healthcare workers during the COVID-19 pandemic in the United States.

In 2020, healthcare workers faced the COVID-19 pandemic amidst other salient sociopolitical stressors. This study, therefore, set out to examine associations between personal, work-related and contextual factors and three outcomes - stress, burnout and turnover intention - at a critical juncture in the pandemic. In December 2020, we recruited a broad array of healthcare workers (n=985) in a public safety net healthcare system serving socially and economically marginalised communities in the Southwest region of the United States using a cross-sectional online survey. The results indicated that more health problems were associated with higher stress and burnout symptoms. While seeking emotional support and using drugs or alcohol to cope were associated with higher stress, a positive social outlook was associated with lower stress. Lower quality of work-life was associated with higher burnout symptoms and turnover intention. Negative effects of the pandemic on wellbeing and higher number of COVID-19-related concerns were associated with higher stress and burnout symptoms. Contrary to the original hypotheses, self-care was not associated with any of the three outcomes, and effects of the political climate and issues of racism on wellbeing were not associated with stress, burnout or turnover intention. However, identifying as a Person of Colour was associated with higher stress, as well as lower burnout. The findings on worker health, social outlook, quality of work-life and race/ethnicity, in particular, suggest a critical need for healthcare systems to address the wellbeing of workers through equitable organisational policy and practice.

Applying a socioecological framework to chronic disease management: implications for social informatics interventions in safety-net healthcare settings.

ObjectiveVulnerable populations face numerous barriers in managing chronic disease(s). As healthcare systems work toward integrating social risk factors into electronic health records and healthcare delivery, we need better understanding of the interrelated nature of social needs within patients’ everyday lives to inform effective informatics interventions to advance health equity.Materials and MethodsWe conducted in-depth interviews, participant-led neighborhood tours, and clinic visit observations involving 10 patients with diabetes in underserved San Francisco neighborhoods and 10 community leaders serving those neighborhoods. We coded health barriers and facilitators using a socioecological framework. We also linked these qualitative data with early persona development, focusing on patients’ experiences in these communities and within the healthcare system, as a starting place for our future informatics design.ResultsWe identified social risk and protective factors across almost every socioecological domain and level—from physical disability to household context to neighborhood environment. We then detailed the complex interplay across domains and levels within two critical aspects of patients’ lives: housing and food. Finally, from these data we generated 3 personas that capture the intersectional nature of these determinants.ConclusionDrawing from different disciplines, our study provides a socioecological approach to understanding health promotion for patients with chronic disease in a safety-net healthcare system, using multiple methodologies. Future digital health research should center the lived experiences of marginalized patients to effectively design and implement informatics solutions for this audience.

Stimulant use for self-management of pain among safety-net patients with chronic non-cancer pain

Background Chronic pain affects one-fifth of US adults. Reductions in opioid prescribing have been associated with increased non-prescription opioid use and, chronologically, increased stimulant (methamphetamine and cocaine) use. While non-prescription opioid use is commonly attributed to pain self-management, the role of stimulants in managing pain is unclear. Methods We analyzed baseline data from a longitudinal study of patients with chronic non-cancer pain in an urban safety-net healthcare system who had been prescribed an opioid for ≥3 of the last 12 months, and had a history of non-prescription opioid, cocaine, or amphetamine use (N = 300). We estimated the prevalence and identified correlates of stimulant use to treat pain among a subgroup of patients who reported past-year stimulant use (N = 105). Data sources included computer-assisted questionnaire (demographics, substance use, pain), clinical exam and procedures (pain, pain tolerance), and chart abstraction (opioid prescriptions). We conducted bivariate analyses to assess associations between demographics, pain characteristics, non-opioid therapies, substance use, opioid prescriptions, and self-reported symptoms, with reporting using stimulants to treat pain. Demographic variables and those with significant bivariate associations were included in a multivariable logistic regression model. Results Fifty-two percent of participants with past-year stimulant use reported using stimulants in the past year to treat pain. Participants who used stimulants for pain reported slightly higher average pain in the past 3 months (median of 8 (IQR: 6–8) vs 7 (7–9) out of 10, p = 0.049). In the multivariable analysis, female gender (AOR= 3.20, 95% CI: 1.06–9.63, p = 0.039) and higher score on the Douleur Neuropathique 4 neuropathic pain questionnaire (AOR = 1.34, 95% CI: 1.05–1.70, p = 0.017) were associated with past-year stimulant use to treat pain. Conclusion Stimulants may be used for pain self-management, particularly for neuropathic pain and among women. Our findings suggest an underexplored motivation for stimulant use in an era of reduced access to prescribed opioids.

618. Characteristics and Outcomes of an Outpatient Infectious Diseases E-consult Program at a County Safety-Net Healthcare System

BackgroundSafety-net healthcare systems often have significant demands for specialty care due to large patient volumes. Infectious Disease (ID) e-consults have the capability to relieve some of this burden by presenting providers with an alternative to face-to-face ID referrals that also lessens financial, travel, and time constraints on patients. Such a system offers the prospect of increasing access to ID care for patients in limited resource settings.MethodsWe performed a retrospective review describing characteristics and outcomes of all outpatient ID e-consults at Parkland Health and Hospital System in Dallas, Texas from March 2018 – February 2021. ResultsIn the study period, 725 e-consults were completed. All e-consults were answered within 72 hours per hospital policy. The most common e-consult topics were 135 (19%) tuberculosis (TB), 116 (16%) syphilis, 97 (13%) respiratory and 79 (11%) musculoskeletal (Figure 1). Nearly two-thirds of the e-consults 456 (63%) came from primary care providers (PCPs). The remainder came from specialists with the most common referring specialties being GI 55 (8%), Hematology/Oncology 36 (5%), Rheumatology 28 (4%) Neurology 27 (4%), and Dermatology 22 (3%) (Figure 2). The majority of e-consults 569 (78%) were resolved without a face-to-face visit. Figure 1. Number of E-consults over Time, by Topic Figure 2. E-consult Topics by Referring Specialty ConclusionImplementation of an outpatient ID e-consult program at a large safety-net healthcare system was an effective means of providing timely input on common ID topics, such as latent TB and interpretation of syphilis serologies, without formal clinic visits. E-consults were able to service a range of providers including PCPs and a variety of specialties, and most e-consults were completed without a clinic visit.Disclosures All Authors: No reported disclosures

Barriers and Facilitators of Implementing a Clinic-Integrated Food Prescription Plus Culinary Medicine Program in a Low-Income Food Insecure Population: A Qualitative Study

BackgroundFood prescription and culinary medicine programs are gaining popularity as tools for decreasing food insecurity, increasing personal agency, promoting healthy eating, and reducing the risk of chronic diseases. However, there is a gap in understanding of how health care professionals can deliver evidence-based how-to nutrition information that is tailored for culturally diverse, low-income populations. ObjectiveTo understand the barriers and facilitators for healthy eating among a low-income, diverse population with diabetes, and the gaps in knowledge and training needed for registered dietitian nutritionists (RDN) to address patient barriers when implementing a food prescription and CM program in a healthcare setting. DesignA series of nine focus groups were conducted: six focus groups with patients with diabetes (n = 40) (three in English and three in Spanish) and three focus groups with RDN employees (n = 17). Participants/settingA convenience sample of 40 low-income food insecure patients with diabetes receiving care at a diverse, integrated, safety net health care system in an urban setting in Texas and convenience sample of 17 RDN employees. Statistical analysisAll focus group transcripts were examined by independent reviewers and blind catalogued and organized into common themes and subthemes based on constant comparative methodology. Investigator group consensus was reached on emergent themes and subthemes for the respective focus groups. ResultsPatients reported frustration with mixed dietary messages from different health care providers, lack of culturally inclusive recommendations, and a desire for skills to prepare tasty and healthy food. RDNs desired more training and education in cultural humility, culinary nutrition skills, and behavioral change theory. ConclusionsOur study describes how cultural humility, practical culinary nutrition skills, and consistent and coordinated messaging can help to improve patient nutrition care. Before implementation of a food prescription and culinary medicine program, efforts should promote training of RDN staff in culinary nutrition and related areas to increase acceptability and adherence of the program for patients.

Colorectal Cancer Screening and Yield in a Mailed Outreach Program in a Safety-Net Healthcare System.

BackgroundScreening with fecal immunochemical testing (FIT) reduces colorectal cancer mortality; however, screening remains low in underserved populations. Mailed outreach, including an invitation letter, FIT, and test instructions, is an evidence-based strategy to improve screening.AimsTo examine screening completion and yield in a mailed outreach program in a safety-net healthcare system.MethodsWe identified and mailed outreach invitations to patients due for screening in a large safety-net system between September 1, 2018, and August 31, 2019. We examined: (1) screening completion, the proportion of patients completing FIT or screening colonoscopy within 6 months of the mailed invitation; and (2) timely diagnostic colonoscopy, the proportion of patients completing colonoscopy within 6 months of positive FIT.ResultsWe mailed 14,879 invitations to 13,190 patients. Nearly half (n = 6098, 46.2%) of patients completed screening: 4,896 (80.3%) completed FIT through mailed outreach; 1,114 (18.3%) FIT through usual care; and 88 (1.4%) screening colonoscopy through usual care. Of patients with a positive FIT (n = 289), 50.5% completed diagnostic colonoscopy within 6 months, 10.7% within 6–12 months, and 4.8% after 12 months. A total of 8 cancers and 83 advanced adenomas were detected in the 191 patients completing diagnostic colonoscopy.ConclusionAfter implementing and scaling up mailed outreach in a safety-net system, about half of patients completed screening, and the majority did so through mailed outreach. However, many patients failed to complete diagnostic colonoscopy after positive FIT. Results highlight the importance of adapting mailed outreach programs to local contexts and constraints of healthcare systems, in order to support efforts to improve CRC screening in underserved populations.

Optimizing Cancer Associated Thrombosis (CAT) Risk Assessment Model at a Safety-Net Healthcare System

Introduction: Cancer associated thrombosis is a preventable complication that impacts the quality of life of patients with cancer. The Khorana score (KS) is the most widely used risk assessment model (RAM) to predict venous thromboembolism (VTE) in ambulatory patients undergoing chemotherapy. Potential limitations of the score include modest discrimination and small proportion of patients in the highest risk subgroup. We aimed to examine if a clinical informatics approach incorporating race/ethnicity, cancer staging, type of systemic therapy, and other known VTE risk factors from the electronic health record (EHR) can improve the RAM.Methods: We performed a retrospective cohort study at Harris Health System (HHS), a safety-net healthcare system that provides care for underserved minorities and uninsured patients in Houston. We created an integrated database that linked consecutive patients with newly diagnosed invasive cancer in the cancer registry with structured data from EPIC Caboodle database 2011-2020. Inclusion/exclusion criteria are shown in Figure 1. We followed patients from time of initial systemic therapy to time of first VTE, death, or loss of follow-up. VTE was defined as radiologically confirmed pulmonary embolism (PE), proximal or distal lower extremity deep vein thrombosis (LE-DVT), catheter-related DVT (CR-DVT), or splanchnic vein thrombosis (SVT) in inpatient or outpatient setting. We used acute, chronic or historical VTE ICD9/ICD10 facility billing codes to assess for potential events and confirmed incident and recurrent events through medical record review. We used multivariable Cox regression to assess potential risk predictors. The model was built iteratively to expand upon the KS. Kaplan Meier failures curves were used to estimate the VTE incidence. C statistic was assessed with binary outcomes at 3- and 6-month.Results: A total of 4,546 patients with newly diagnosed cancer receiving 1 st line systemic therapy met the inclusion/exclusion criteria. Relevant demographics showed a median age of 54 (IQR 46-61), 57% female, 50% Hispanic, 27% Black, and 75% uninsured. Most common cancer types included breast (17%), colorectal (13%), lung (10%), and non-Hodgkin lymphoma (8%); 32% of patients had metastatic disease. First-line systemic therapy included 89% cytotoxic chemotherapy, 9% small molecule targeted +/- endocrine therapy, and 2% PD-1/PD-L1 immunotherapy. Only 1% had remote VTE history after excluding 317 patients already on therapeutic anticoagulation.Incident VTE occurred in 477 patients during a median follow-up of 11.3 months. There were 229 PE +/- other, 140 LE-DVT, 94 CR-DVT, and 14 SVT. In addition to the KS covariates (Table 1), recent cancer diagnosis (≤ 1 month) (HR 1.43, 1.16-1.76), metastatic disease (HR 1.46, 1.20-1.76), recent hospitalization (≤ 3 month) (HR 1.54, 1.24-1.90) were also associated with higher risk of VTE, whereas Hispanic ethnicity (HR 0.65, 0.51-0.84) and Asian race (HR 0.30, 0.16-0.54) were associated with a lower risk. Other appreciable predictors included targeted vs. chemotherapy (HR 0.75, 0.51-1.10) and history of PE/LE-DVT (HR 1.74, 0.93-3.27). Immunotherapy (vs. chemotherapy) and black (vs. white) were not associated with VTE.Figure 2 shows the comparison of VTE incidence using the 2 RAMs. Original KS RAM had c statistic of 0.66 and 0.62 at 3- and 6-month, respectively (Table 2). The highest risk group (3+) included 16% of patients (n=723) and 23% of all VTE (n=108). The modified RAM had better discrimination with c statistic of 0.71 and 0.67 at 3- and 6-month, respectively. The highest risk group (3+) included 33% of patients (n=1509) and 51% of all VTE (n=244).Conclusions: The KS performed reasonably well in a large safety-net healthcare system with predominantly uninsured patients with advanced cancer initiating systemic therapy. Nonetheless, simple structured data elements from the EHR such as race/ethnicity, staging, therapy type, and recent hospitalization improved the performance of the KS-based RAM and doubled the number of patients in the high-risk stratum and the number of preventable VTE. An integrated clinical informatics approach adapted to the local population can improve outcomes by identifying patients most appropriate for ambulatory thromboprophylaxis. [Display omitted] DisclosuresCarrier: Sanofi: Honoraria; Leo Pharma: Honoraria, Research Funding; Servier: Honoraria; Pfizer: Honoraria, Research Funding; Bayer: Honoraria; BMS: Honoraria, Research Funding.

Impact of Race/Ethnicity on Cancer Associated Thrombosis Among Underserved Patients with Cancer

Introduction: Cancer-associated thrombosis (CAT) is common among patients with cancer. Risk factors for CAT include type of malignancy, advanced stage, and chemotherapy treatment, but the association of CAT with race and ethnicity remains controversial. Identifying the incidence of CAT among populations susceptible to inequalities in healthcare delivery may help delineate preventive strategies.Methods: We performed a retrospective cohort study at Harris Health System (HHS), a safety-net healthcare system that provides care for underserved minorities and uninsured patients in Houston, TX. We created an integrated database that linked consecutive patients with newly diagnosed invasive cancer with structured electronic health record (EHR) data from 2011-2020 (Figure 1). We followed patients from time of cancer diagnosis to time of first VTE, death, or loss of follow-up. VTE was defined as radiologically confirmed pulmonary embolism (PE), lower extremity deep vein thrombosis (LE-DVT), catheter-related DVT (CR-DVT), or splanchnic vein thrombosis in either inpatient or outpatient setting. We used VTE ICD9/ICD10 billing codes to assess for potential events and confirmed incident, recurrent, and historical events through medical record review. VTE occurring within 30 days prior to cancer diagnosis were considered as CAT at diagnosis.Incidence rates were assessed per 100 person-year (py) within 1 year of diagnosis and stratified by race/ethnicity, cancer type, and cancer stage. Cumulative incidence of VTE was assessed through competing risk method with death as the competing cause. Multivariable Fine-Gray competing risk models were performed to determine the effect of race/ethnicity on the risk of CAT, adjusted for age, sex, body mass index, insurance, cancer site, stage, systemic therapy, recent hospitalization, and prior history of VTE.Results: A total of 9,353 cancer patients were included in the study, where 49.3% were Hispanics, 27.6% were Non-Hispanic Blacks (NHB), 15.5% were Non-Hispanic Whites (NHW), and 7.6% were Asian/Pacific Islander (PI). Most patients (74.7%) were uninsured, 35.8% were obese, 19% had recent hospitalization, and 31.9% had stage IV disease. Overall, 832 developed CAT within 1 year, including 49.4% PE, 28.1% LE-DVT, and 17.1% CR-DVT. The median onset was 69 days (IQR 20-154), but a significant proportion (n=92) was diagnosed in the month before diagnosis. The incidence of CAT was 7.3% at 6 months and 9.6% at 1 year. The overall incidence rate was 11/100 py with a similar trend from 2011 to 2020.Figure 2 shows the variation in incidence rates for different cancer types across stages. The rate increased 2- to 10-fold from stage I to IV, reaching >40/100 py among patients with pancreatic and upper gastrointestinal cancers. Figure 3 shows the impact of race/ethnicity on the incidence of CAT, with 9.3% and 8.4% for NHW and NHB at 6 months, compared to 6.5% and 3.8% for Hispanics and Asian/PI, respectively. In the adjusted multivariable analysis, the risk of CAT remained lower in Hispanics vs. NHW (SHR 0.80 [0.65-0.97]) and Asian/PI vs. NHW (SHR 0.48 [0.32-0.73]). There was no difference in NHB vs. NHW (SHR 1.04 [0.84-1.27]). Other important covariates included history of VTE (SHR 2.29 [1.32-3.97]) and prolonged hospitalization (SHR 1.53 [CI 1.31-1.80]) in addition to staging and cancer types. Compared to no systemic therapy or adjuvant endocrine only, there was a higher incidence of VTE in patients receiving chemotherapy (SHR 1.72 [1.39-2.15]), targeted therapy (SHR 1.58 [1.12-2.23]), immunotherapy (SHR 2.13 [1.30-3.48], but not non-adjuvant endocrine therapy (SHR 0.91 [0.40-2.08]).Conclusion: In the current cancer registry-linked EHR cohort with a high proportion of Hispanics and Blacks from a large safety-net healthcare system in the US, we have observed a high incidence rate of CAT at 11/100 py. This rate is much higher than 3.9-5.8/100 py reported in contemporary cancer registries in Europe (PMID 27709226, 33171494), and likely reflects advanced disease and comorbidity at the time of diagnosis due to delayed care from the lack of health insurance. Despite adjusting for patient-, cancer-, and treatment-specific confounders, we found that Hispanics and Asian/PI had ~20% and ~50% lower rate of VTE, respectively, compared to NHW or NHB. This racial/ethnicity difference should be considered in future risk assessment models for CAT. [Display omitted] DisclosuresCarrier: Sanofi Aventis: Honoraria, Membership on an entity's Board of Directors or advisory committees; LEO Pharma: Honoraria, Membership on an entity's Board of Directors or advisory committees; Bayer: Honoraria, Membership on an entity's Board of Directors or advisory committees; Bristol Myers Squibb: Honoraria; Pfizer: Honoraria, Membership on an entity's Board of Directors or advisory committees; Aspen: Membership on an entity's Board of Directors or advisory committees; Boehringer Ingelheim: Honoraria; Servier: Honoraria.

HIV Testing Strategies for Health Departments to End the Epidemic in the U.S.

An important goal of the Ending the HIV Epidemic in the U.S. initiative is the timely diagnosis of all people with HIV as early as possible after infection. To end the HIV epidemic, health departments were encouraged to propose new and innovative HIV testing strategies and improve the reach of existing programs. These activities were divided into 3 core strategies: expansion of routine screening in healthcare settings, locally tailored HIV testing initiatives in nonhealthcare settings, and specific efforts to increase the frequency of testing for individuals with increased potential for acquiring HIV. Because HIV testing is such a crucial part of the core activities of the Centers for Disease Control and Prevention's HIV prevention programs, there are many examples of evidence-based programs and best practices for HIV testing in both clinical and nonclinical settings. This article reviews the evidence base for these strategies and some of the activities proposed under the Diagnose pillar to achieve the goal of diagnosing all HIV infections as early as possible. All other Ending the HIV Epidemic in the U.S. activities start with an awareness of HIV status, which is actually the indicator for which most health departments are closest to the proposed targets. There are both proven and emerging approaches to increasing HIV screening and increasing the frequency of HIV screening available. The Ending the HIV Epidemic in the U.S. initiative provides the motivation, the resources, and a coordinated plan to bring them to scale.

Disparities and Temporal Trends in COVID-19 Exposures and Mitigating Behaviors Among Black and Hispanic Adults in an Urban Setting

COVID-19, caused by SARS-CoV-2 virus, has disproportionately affected Black and Hispanic communities in the US, which can be attributed to social factors including inconsistent public health messaging and suboptimal adoption of prevention efforts. To identify behaviors and evaluate trends in COVID-19-mitigating practices in a predominantly Black and Hispanic population, to identify differences in practices by self-reported ethnicity, and to evaluate whether federal emergency financial assistance was associated with SARS-CoV-2 acquisition. This survey study was conducted by telephone from July 1 through August 30, 2020, on a random sample of adults who underwent SARS-CoV-2 testing at a safety-net health care system in Chicago during the surge in COVID-19 cases in the spring of 2020. Behaviors and receipt of a stimulus check were compared between participants testing positive and negative for SARS-CoV-2. Differences in behaviors and temporal trends were assessed by race and ethnicity. SARS-CoV-2 infection was assessed using nasopharyngeal quantitative reverse transcriptase-polymerase chain reaction testing. Mitigating behaviors and federal emergency financial assistance were assessed by survey. Race and ethnicity data were collected from electronic health records. Of 750 randomly sampled individuals, 314 (41.9%) consented to participate (169 [53.8%] women). Of those, 159 (51%) self-reported as Hispanic and 155 (49%) as non-Hispanic (120 [38.2%] Black), of whom 133 (84%) and 76 (49%) tested positive for SARS-CoV-2, respectively. For all participants, consistent mask use (public transport: adjusted odds ratio [aOR], 0.00; 95% CI, 0.00-0.34; social gatherings: aOR, 0.10; 95% CI, 0.00-0.50; running errands: aOR, 0.18; 95% CI, 0.07-0.42; at work: aOR, 0.23; 95% CI, 0.07-0.79) and hand sanitizer use (aOR, 0.26; 95% CI, 0.13-0.52) were associated with lower odds of infection. During 3 sampled weeks, mitigation practices were less frequent among Hispanic compared with non-Hispanic participants (eg, mask use while running errands: aOR, 0.26; 95% CI, 0.15-0.46). Hispanic participants were at high risk of infection (aOR, 5.52; 95% CI, 4.30-7.08) and more likely to work outside the home (aOR, 2.05; 95% CI, 1.27-3.30) compared with non-Hispanic participants, possibly because of limited receipt of stimulus checks (aOR, 0.03; 95% CI, 0.02-0.07) or unemployment benefits (aOR, 0.36; 95% CI, 0.16-0.74). In this survey study of adults in a large US city, public health messaging improved preventive behaviors over time but lagged among Hispanic participants; messaging tailored to Hispanic communities, especially for mask use, should be prioritized. Hispanic individuals were at higher risk for infection, more often worked outside the home, and were less likely to have received a stimulus check; this suggests larger studies are needed to evaluate the provision of economic support on SARS-CoV-2 transmission dynamics in low-income populations.

A Cross-Sectional Analysis of Barriers Associated With Non-Attendance at a Urology Telehealth Clinic in a Safety-Net Hospital

To analyze the factors associated with non-attendance at a urology telehealth clinic in a large urban safety-net hospital after institutional-mandated transition to telehealth due to COVID-19. We identified all encounters scheduled for telehealth after March 17, 2020 and in the subsequent 8 weeks. Logistic regression was used to identify factors associated with attendance. In total there were 322 telehealth encounters, 228 (70.8%) of which were attended and 94 (29.2%) that were not attended. Racial/ethnic minorities accounted for 175 (77.0%) of attended and 73 (76.7%) of non-attended encounters. On multivariable regression, single/divorced/widowed (odds ratio [OR] 2.36, 95% confidence interval [CI] 1.26-4.43), current substance use disorder (OR 5.33, 95% CI 2.04-13.98), and being scheduled for a new patient appointment (OR 1.81, 95% CI 1.04-3.13) were associated with higher odds of not attending a telehealth encounter. Race/ethnicity, primary language, and country of birth were not associated with odds of attendance. Our findings identify several social factors (social support, substance use) associated with non-attendance at outpatient telehealth urology encounters at an urban safety-net hospital during the early stages of the COVID-19 pandemic. These barriers may have a greater impact specifically within a safety-net healthcare system and will inform equitable provision of urology telehealth programs in the future FUNDING: Goldberg-Benioff Endowed Professorship in Cancer Biology. The sponsors had no involvement with this study.

Improving Timeliness of Antibiotic Administration Using a Provider and Pharmacist Facing Sepsis Early Warning System in the Emergency Department Setting: A Randomized Controlled Quality Improvement Initiative.

Results of pre-post intervention studies of sepsis early warning systems have been mixed, and randomized clinical trials showing efficacy in the emergency department setting are lacking. Additionally, early warning systems can be resource-intensive and may cause unintended consequences such as antibiotic or IV fluid overuse. We assessed the impact of a pharmacist and provider facing sepsis early warning systems on timeliness of antibiotic administration and sepsis-related clinical outcomes in our setting. A randomized, controlled quality improvement initiative. The main emergency department of an academic, safety-net healthcare system from August to December 2019. Adults presenting to the emergency department. Patients were randomized to standard sepsis care or standard care augmented by the display of a sepsis early warning system-triggered flag in the electronic health record combined with electronic health record-based emergency department pharmacist notification. The primary process measure was time to antibiotic administration from arrival. A total of 598 patients were included in the study over a 5-month period (285 in the intervention group and 313 in the standard care group). Time to antibiotic administration from emergency department arrival was shorter in the augmented care group than that in the standard care group (median, 2.3 hr [interquartile range, 1.4-4.7 hr] vs 3.0 hr [interquartile range, 1.6-5.5 hr]; p = 0.039). The hierarchical composite clinical outcome measure of days alive and out of hospital at 28 days was greater in the augmented care group than that in the standard care group (median, 24.1 vs 22.5 d; p = 0.011). Rates of fluid resuscitation and antibiotic utilization did not differ. In this single-center randomized quality improvement initiative, the display of an electronic health record-based sepsis early warning system-triggered flag combined with electronic health record-based pharmacist notification was associated with shorter time to antibiotic administration without an increase in undesirable or potentially harmful clinical interventions.

Trends in prevalence of extended-spectrum beta-lactamase-producing Escherichia coli isolated from patients with community- and healthcare-associated bacteriuria: results from 2014 to 2020 in an urban safety-net healthcare system

BackgroundThe prevalence of infections caused by extended-spectrum beta-lactamase producing Escherichia coli (ESBL-E. coli) is increasing worldwide, but the setting in which this increase is occurring is not well defined. We compared trends and risk factors for ESBL-E. coli bacteriuria in community vs healthcare settings.MethodsWe collected electronic health record data on all patients with E. coli isolated from urine cultures in a safety-net public healthcare system from January 2014 to March 2020. All analyses were stratified by healthcare-onset/associated (bacteriuria diagnosed > 48 h after hospital admission or in an individual hospitalized in the past 90 days or in a skilled nursing facility resident, N = 1277) or community-onset bacteriuria (bacteriuria diagnosed < 48 h after hospital admission or in an individual seen in outpatient clinical settings without a hospitalization in the past 90 days, N = 7751). We estimated marginal trends from logistic regressions to evaluate annual change in prevalence of ESBL-E. coli bacteriuria among all bacteriuria. We evaluated risk factors using logistic regression models.ResultsESBL-E. coli prevalence increased in both community-onset (0.91% per year, 95% CI 0.56%, 1.26%) and healthcare-onset/associated (2.31% per year, CI 1.01%, 3.62%) bacteriuria. In multivariate analyses, age > 65 (RR 1.88, CI 1.17, 3.05), male gender (RR 2.12, CI 1.65, 2.73), and Latinx race/ethnicity (RR 1.52, CI 0.99, 2.33) were associated with community-onset ESBL-E. coli. Only male gender (RR 1.53, CI 1.03, 2.26) was associated with healthcare-onset/associated ESBL-E. coli.ConclusionsESBL-E. coli bacteriuria frequency increased at a faster rate in healthcare-associated settings than in the community between 2014 and 2020. Male gender was associated with ESBL-E. coli bacteriuria in both settings, but additional risks—age > 65 and Latinx race/ethnicity—were observed only in the community.

Perceptions on Barriers and Facilitators to Colonoscopy Completion After Abnormal Fecal Immunochemical Test Results in a Safety Net System

The effectiveness of stool-based colorectal cancer (CRC) screening, including fecal immunochemical tests (FITs), relies on colonoscopy completion among patients with abnormal results, but in safety net systems and federally qualified health centers, in which FIT is frequently used, colonoscopy completion within 1 year of an abnormal result rarely exceeds 50%. Clinician-identified factors in follow-up of abnormal FIT results are understudied and could lead to more effective interventions to address this issue. To describe clinician-identified barriers and facilitators to colonoscopy completion among patients with abnormal FIT results in a safety net health care system. This qualitative study was conducted using semistructured key informant interviews with primary care physicians (PCPs) and staff members in a large safety net health care system in Washington state. Eligible clinicians were recruited through all-staff meetings and clinic medical directors. Interviews were conducted from February to December 2020 through face-to-face interactions or digital meeting platforms. Interview transcripts were analyzed deductively and inductively using a content analysis approach. Data were analyzed from September through December 2020. Barriers and facilitators to colonoscopy completion after an abnormal FIT result were identified by PCPs and staff members. Among 21 participants, there were 10 PCPs and 11 staff members; 20 participants provided demographic information. The median (interquartile range) age was 38.5 (33.0-51.5) years, 17 (85.0%) were women, and 9 participants (45.0%) spent more than 75% of their working time engaging in patient care. All participants identified social determinants of health, organizational factors, and patient cognitive factors as barriers to colonoscopy completion. Participants suggested that existing resources that addressed these factors facilitated colonoscopy completion but were insufficient to meet national follow-up colonoscopy goals. In this qualitative study, responses of interviewed PCPs and staff members suggested that the barriers to colonoscopy completion in a safety net health system may be modifiable. These findings suggest that interventions to improve follow-up of abnormal FIT results should be informed by clinician-identified factors to address multilevel challenges to colonoscopy completion.

Detecting diabetic retinopathy through machine learning on electronic health record data from an urban, safety net healthcare system.

ObjectiveClinical guidelines recommend annual eye examinations to detect diabetic retinopathy (DR) in patients with diabetes. However, timely DR detection remains a problem in medically underserved and under-resourced settings in the United States. Machine learning that identifies patients with latent/undiagnosed DR could help to address this problem.Materials and MethodsUsing electronic health record data from 40 631 unique diabetic patients seen at Los Angeles County Department of Health Services healthcare facilities between January 1, 2015 and December 31, 2017, we compared ten machine learning environments, including five classifier models, for assessing the presence or absence of DR. We also used data from a distinct set of 9300 diabetic patients seen between January 1, 2018 and December 31, 2018 as an external validation set.ResultsFollowing feature subset selection, the classifier with the best AUC on the external validation set was a deep neural network using majority class undersampling, with an AUC of 0.8, the sensitivity of 72.17%, and specificity of 74.2%.DiscussionA deep neural network produced the best AUCs and sensitivity results on the test set and external validation set. Models are intended to be used to screen guideline noncompliant diabetic patients in an urban safety-net setting.ConclusionMachine learning on diabetic patients’ routinely collected clinical data could help clinicians in safety-net settings to identify and target unscreened diabetic patients who potentially have undiagnosed DR.

A Cross-Sectional Study of SARS-CoV-2 Vaccination Among Employees of an Urban Safety-Net Health Care System.

A Cross-Sectional Study of SARS-CoV-2 Vaccination Among Employees of an Urban Safety-Net Health Care System.

829-P: Improving Diabetes Self-Management with a Peer-to-Peer Mentoring Program: Final Results of a Pilot Program at a Safety Net Health Care System

New York City Health + Hospitals (NYC H+H) is the largest safety net healthcare system in the nation, serving more than one million patients each year. Most of NYC H+H’s patients are low-income and people of color, who, in NYC as in the rest of the nation, bear a disproportionate burden of chronic diseases. At NYC H+H, 18% of patients have diagnosed diabetes, and nearly 23,000 (39%) of the 59,000 patients monitored in NYC H+H’s diabetes registry have poorly controlled diabetes (A1c ≥ 8). In an effort to support these patients’ diabetes management between their regular visits, NYC H+H launched a pilot program of InquisitHealth, a one-on-one peer mentoring service that matches patients with mentors who have controlled their own diabetes and received training to become peer coaches. Several small, randomized, controlled trials have demonstrated the program’s effectiveness in reducing A1c, but there is limited evidence of its effectiveness in a real-word setting. The key aims of this pilot program were to: 1) bridge the gap between patients and providers outside of clinic visits, 2) increase patient education on and support for diabetes self-management, and 3) improve patients’ diabetes outcomes as evidenced by reductions in A1c. The pilot program enrolled 336 patients at two NYC H+H facilities for a twelve-month intervention. Among the first 180 patients enrolled at one large acute care hospital, the mean reduction in A1c from baseline to six months following enrollment was 1.08. As compared to propensity score-matched control patients at a similar non-participating hospital in the NYC H+H network, participants experienced a greater reduction in A1c of 0.61 (-1.08 vs. -0.47, p&amp;lt;0.01). Results from the 12-month final analysis of all participants will be presented. Disclosure R. Kalyanaraman marcello: None. J. Dolle: None. S. Kaur: None. S. R. Patterson: None. N. Davis: None. Funding National Institutes of Health (UL1TR001445)

Real-World Data on the Impact of COVID-19 on Endoscopic Procedural Delays.

INTRODUCTION:The initial surge of the coronavirus disease 2019 (COVID-19) pandemic prompted national recommendations to delay nonurgent endoscopic procedures. The objective of this study was to provide real-world data on the impact of COVID-19 on endoscopic procedures in a safety-net healthcare system and cancer center affiliated with a tertiary academic center.METHODS:This retrospective cohort study used a combination of electronic health record data and a prospective data tool created to track endoscopy procedures throughout COVID-19 to describe patient and procedural characteristics of endoscopic procedures delayed during the initial COVID-19 surge.RESULTS:Of the 480 patients identified, the median age was 57 years (interquartile range 46–66), 55% (n = 262) were male, and 59% self-identified as white. Colonoscopy was the most common type of delayed procedure (49%), followed by combined esophagogastroduodenoscopy (EGD) and colonoscopy (22%), and EGD alone (20%). Colorectal cancer screening was the most common indication for delayed colonoscopy (35%), and evaluation of suspected bleeding (30%) was the most common indication for delayed combined EGD and colonoscopy. To date, 46% (223/480) of delayed cases have been completed with 12 colorectal, pancreatic, and stomach cancers diagnosed. Sociodemographic factors, procedure type, and sedation type were not significantly associated with endoscopy completion. The median time to endoscopy after delayed procedure was 88 days (interquartile range 63–119) with no differences by procedure type.DISCUSSION:To minimize potential losses to follow-up, delayed, or missed diagnoses and to reduce progression of gastrointestinal diseases, all efforts should be used to ensure follow-up in those whose endoscopic procedures were delayed because of COVID-19.

Pre-screening as a tool for increasing oncology trial enrollment.

1577 Background: Recruiting underserved patients onto therapeutic oncology trials is imperative in light of cancer disparities. Many institutions have increased enrollment with data mining tools that match patients to clinical trials, but such expensive tools are unavailable in a public safety net healthcare system. Methods: The NYU Perlmutter Cancer Center Clinical Trials Office implemented a quality improvement program aimed to increase therapeutic trial enrollment at Bellevue Cancer Center (BCC), an affiliated public hospital. The initiative utilized one employee to manually pre-screen patients via the EMR, cancer registries, and conferences. The program aimed to identify eligible patients for therapeutic trials and those subsequently enrolled. Results: During the two years preceding the pre-screening program, 31 patients were enrolled onto therapeutic clinical trials at BCC. For a period of two years (7/2017-7/2019) following the initiation of this program, 255 patients were identified, of which 143 (56.1%) enrolled onto trials. Of those enrolled, 55 were referred to NYU for trials not open at BCC. Among the 143 enrolled patients, 64% were female, 56% were non-white, and 57% did not speak English (spanning 16 languages). The median age was 55, and the top three disease groups were breast, GI, and thoracic. 83% of our patients reside in low-income areas, with 62% in both low income and Health Professional Shortage Areas (HPSA). Conclusions: Dedicating one employee to screening led to a 4.6 fold increase in accruals, successfully augmenting therapeutic trial enrollment in a healthcare setting with scarce resources, providing broader access to clinical trials for underserved populations.[Table: see text]

Informatics response to address the COVID-19 pandemic in a safety net healthcare system.

In service of particularly vulnerable populations, safety net healthcare systems must nimbly leverage health information technology (IT), including electronic health records (EHRs), to coordinate the medical and public health response to the novel coronavirus (COVID-19). Six months after the San Francisco Department of Public Health implemented a new EHR across its hospitals and citywide clinics, California declared a state of emergency in response to COVID-19. This paper describes how the IT and informatics teams supported San Francisco Department of Public Health’s goals of expanding the safety net healthcare system capacity, meeting the needs of specific vulnerable populations, increasing equity in COVID-19 testing access, and expanding public health analytics and research capacity. Key enabling factors included critical partnerships with operational leaders, early identification of priorities, a clear governance structure, agility in the face of rapidly changing circumstances, and a commitment to vulnerable populations.