Safety Net Clinics Research Articles

Abstract P33: Disparities in breast cancer screening and the impact of COVID-19 in Houston, Texas

Abstract Background: The third most populous county in the United States, the Harris Country, has some of the worst outcomes of breast cancer in the country. In this county, 27% of mammogram eligible women, ages 40-64 are uninsured, compared to about 11% the national average. Harris Health (HH) is a state funded network of SafetyNet hospitals and clinics that provide comprehensive care, including breast cancer screening to the uninsured and underinsured residents of the Harris county. The population who receives care at the HH sites is majority of racial and ethnic minorities and generally lower socioeconomic status than the county average. The number of screening mammograms abruptly dropped by 94% across the United States following the declaration of the COVID-19 national emergency. Currently there is no data on how breast cancer screening rates have been affected during the pandemic across Houston, and specifically in Harris County where health disparities gap already existed. We hypothesized that the population across HH sites have lower rates of screening mammograms following the COVID-19 pandemic declaration compared to the rest of the population across Houston who visit Baylor St. Luke’s Health- Texas (SLH) catchment area. Methods: We retrieved the monthly gross number of screening mammograms performed at all HH sites and SLH sites from February to October for both 2019 and 2020 on the PENRAD database. We then used the Independent-Samples Mann-Whitney U Test to assess differences in the distribution of percent change in mammogram between HH sites and SLH sites. First, we compared the total mammogram numbers for February to October (2020 vs. 2019). We then compared only the numbers for each month (e.g., May 2020 vs. May 2019). Beyond basic descriptive counts, statistical analysis was limited to mammogram among women ages 40 – 69. Statistical significance was set at p<0.05. Results: Overall, across both health care systems, HH and SLH, and for all ages, there was a drastic drop (53.96%) in screening mammogram between 2019 (N=32,968) and 2020 (N=17,788). For women ages 40 – 69, the drop in screening mammogram at SLH sites during 2020 (February to October) was significantly less than the drop at HH sites during the same period. When screening mammogram was compared between 2019 and 2020 for each month separately, it was only in May and September that the % change in mammogram were similar for SLH and HH. For the remaining months, SLH site had significantly less drop in screening mammograms. Conclusion: The population that normally receives care at HH site had significantly lower rates of screening mammograms following the pandemic compared to the population who receives care at SLH sites. This will likely widen the already present disparity gap in breast cancer outcomes for the city of Houston. Citation Format: Maryam Nemati Shafaee, Luke Gilman, Abiodun Oluyomi, Melissa Bondy, Tamara Ortiz-Perez, Chris Amos, Mothaffar Rimawi. Disparities in breast cancer screening and the impact of COVID-19 in Houston, Texas [abstract]. In: Proceedings of the AACR Virtual Meeting: COVID-19 and Cancer; 2021 Feb 3-5. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(6_Suppl):Abstract nr P33.

Racial Disparities in Diabetes Technology Use and Outcomes in Type 1 Diabetes in a Safety-Net Hospital.

Limited data exist regarding diabetes technology use among adults with type 1 diabetes (T1D) in urban racially/ethnically diverse safety-net hospitals. We examined racial/ethnic differences in the use of continuous glucose monitor (CGM) and continuous subcutaneous insulin infusion (CSII) in this setting. A retrospective review of 227 patients ≥ 18 years of age with T1D seen in an urban, safety-net endocrinology clinic during 2016-2017 was completed (mean age: 39; 80% English-speaking; 50% had public insurance). Diabetes technology use, defined as either CGM or CSII or both CGM and CSII, and clinical outcomes were examined by race/ethnicity. Overall, 30% used CGM and 26% used CSII. After adjusting for age, language, insurance, and annual income, diabetes technology use in non-White patients was significantly lower than in White patients, predominantly lower in Black (aOR 0.25 [95% CI 0.11-0.56]) and patients identified as other race/ethnicity (aOR 0.30 [95% CI 0.11-0.77]). At the highest household income level (≥$75,000/y), Black and Hispanic individuals were significantly less likely than White individuals to use diabetes technology (P < .0007). Mean hemoglobin A1c (HbA1c) was lower in patients using any diabetes technology compared with patients using no technology (P < .0001). Use of CGM and CSII together was associated with the lowest HbA1c across all racial/ethnic groups. Racial/ethnic disparities in diabetes technology use and glycemic control were observed even after adjusting for sociodemographic factors. Further research should explore barriers to accessing diabetes technology in non-White populations.

Accessibility of federally funded family planning services in South Carolina and Alabama.

This study operationalized the five dimensions of health care access in the context of contraceptive service provision and used this framework to examine access to contraceptive care at health department (HD) (Title X funded) and federally qualified health center (FQHC) (primarily non-Title X funded) clinics in South Carolina and Alabama. A cross-sectional survey was conducted in 2017/18 that assessed clinic-level characteristics, policies, and practices related to contraceptive provision. Provision of different contraceptive methods was examined between clinic types. Survey items were mapped to the dimensions of access and internal consistency for each scale was tested with Cronbach's alpha. Scores of access were developed and differences by clinic type were evaluated with an independent t-test. The overall response rate was 68.3% and the sample included 235 clinics. HDs (96.9%) were significantly more likely to provide IUDs and/or Impants on-site than FQHCs (37.4%) (P<0.0001). Scales with the highest consistency were Availability: Clinical Policy (24 items) (alpha=0.892) and Acceptability (43 items) (alpha=0.834). HDs had higher access scores than FQHCs for the Availability: Clinical Policy scale (0.58, 95% CL 0.55, 0.61) vs (0.29, 95% CL 0.25, 0.33) and Affordability: Administrative Policy scale (0.86, 95% CL 0.83, 0.90) vs (0.47, 95% CL 0.41, 0.53). FQHCs had higher access scores than HDs for Affordability: Insurance Policy (0.78, 95% CL 0.72, 0.84) vs (0.56, 95% CL 0.53, 0.59). These findings highlight strengths and gaps in contraceptive care access. Future studies must examine the impact of each dimension of access on clinic-level contraceptive utilization.

National profile of the growing population of older adults who access community health centers.

Community health centers (CHCs) are federally funded safety-net clinics that provide care to low income and medically underserved persons. The proportion of CHC patients aged ≥65 doubled in the last ten years, yet little is known about this population. We aim to describe the demographic and clinical characteristics of the older adult CHC population. Cross sectional analysis. The nationally representative 2014 Health Center Patient Survey. CHC patients ≥55 years. We used descriptive statistics to characterize older adults across demographic and clinical variables. To determine differences by age, we stratified into three groups (55-64, 65-74, 75+ years). We used t-tests and chi-squared to calculate p values and survey weights to make national estimates. We included 1875 older adults ≥55 years, representing over 4.2 million people. Older adults were mostly aged 55-64 (60%), female (51%), and white (60%). The majority (73%) had Medicare or Medicaid and 47% reported fair or poor health. Regardless of age, older adults had an average of three chronic conditions and 0.6 impairments in activities of daily living (ADL). Healthcare utilization was not significantly different across age groups with most taking ≥5 prescription medications (54%) and one in five reporting ≥2 emergency department visits or ≥1 hospitalization in the last year. Adults 55-64 who attend CHCs have similar disease burden as adults ≥65. As the population of older adults who access CHCs grow, our findings highlight the opportunity to enhance focus on key principles of geriatric medicine, such as measurement of functional impairment for those who are <65 while also addressing underlying health disparities.

How a Medication for Opioid Use Disorder Curriculum Translates into Experiences and Internal Medicine Residents’ Understanding of Patients with Opioid Use Disorder

Problem The number of people with an Opioid Use Disorder (OUD) continues to outpace access to associated medication. Ninety-six percent of states report higher rates of OUD than access to medications, and, despite being the standard of care, only 3% of physicians currently prescribe medication for opioid use disorder (MOUD). Prior studies have shown that decreasing barriers, such as a lack of knowledge about MOUD, increased physicians’ willingness to prescribe. However, most internal medicine residency programs do not have a required addiction curriculum. As a result, we created a curriculum and conducted qualitative interviews with residents to better understand experiences with the curriculum. Intervention In an effort to overcome physician-centered barriers associated with prescribing MOUD, we developed and implemented a week-long curriculum, Addiction Week, for second and third year Internal Medicine Residents at Indiana University School of Medicine in a safety-net clinic. The curriculum included the following: didactics on substance use disorder (SUD), including OUD and alcohol use disorder, and MOUD (mostly buprenorphine), and mostly web-based, peer-reviewed and guideline based readings about addiction, direct observation of addiction counselors, direct discussion with people receiving MOUD, observation of a group therapy session, informal discussion with providers who prescribe MOUD, and, for some residents, observation of a physician prescribing MOUD. After completing the curriculum, the residents participated in an hour long audio-recorded interview to better understand their experiences with the curriculum. Context This study was completed at a residency program where residents were not previously exposed to outpatient MOUD prescribing. Due to limited availability of faculty treating patients with MOUD, residents spent the majority of their time shadowing a social worker. Impact Residents described gaining a deeper understanding of OUD by having the opportunity to interact with patients in a stable outpatient setting, which for many led to increased confidence and willingness to prescribe MOUD for people with OUD. Lessons learned The greater understanding of addiction and willingness to prescribe MOUD described by residents in this study indicate that this type of curriculum may be a promising way to increase MOUD prescribing. Further studies are needed to evaluate whether this intervention can change prescribing behaviors. Supplemental data for this article is available online at https://doi.org/10.1080/10401334.2021.1897597.

Patient Satisfaction in a Safety Net Urology Clinic

Patient satisfaction has become an important metric in medicine. In this study, we aim to identify modifiable factors contributing toward poor satisfaction amongst patients in a safety net urology clinic. We hypothesize that wait times and long distances traveled to clinic will have the largest contribution toward negative patient experiences. We conducted a cross-sectional study of adult patients presenting to the urology clinic at Zuckerberg San Francisco General Hospital. Participants completed a survey using the RAND Patient Satisfaction Questionnaire-18 after their clinic visit. Associations among demographic factors, wait times, use of a translator and satisfaction were examined. Two hundred patients, 19 to 90 years old (median age 62), were enrolled. Overall, 65% of patients were satisfied with their experience in our clinic. Our results demonstrated a significant association among race, language and overall satisfaction scores (p=0.009 and p=0.003, respectively). Multivariable analysis showed that those who waited more than an hour to see a physician were less satisfied compared with those who waited less than 15 minutes (OR 0.25, 95% CI 0.08-0.74). Similarly, those who used a translator were less satisfied than those who did not use a translator (OR 0.29, 95% CI 0.08-1.01). In a safety net urology clinic, language differences and wait times may contribute to poor patient satisfaction. Future efforts toward improving language interpreter functionality and decreasing patient wait times will likely improve patient satisfaction.

Addressing COVID-19 Using a Public Health Approach: Perspectives From the Cancer Prevention and Control Research Network

Addressing COVID-19 Using a Public Health Approach: Perspectives From the Cancer Prevention and Control Research Network

Genomic landscape of advanced prostate cancer in racial minority populations: Real-world experience in a safety-net hospital oncology clinic.

14 Background: The largest US cancer health disparity exists in prostate cancer (PC), with African American (AA) men having: ~1.6-1.8-fold higher risk of developing PC; younger age and more advanced stage at diagnosis; increased risk of recurrence after radical prostatectomy; and up to 2.5-fold higher mortality rate relative to men of other ancestries. Access to healthcare and other socioeconomic and environmental factors contribute to the disparity in clinical outcomes. However, genetic factors may also be involved, and their role and prevalence need to be better defined, especially in real-world clinical settings, as the high cost of next-generation sequencing (NGS) may have resulted in underrepresentation of uninsured and minority patients in prior studies. Methods: We retrospectively analyzed NGS data obtained via Tempus|xT tissue assay (DNA sequencing of 648 genes in tumor and matched normal samples at 500x depth) and/or Tempus|xF liquid biopsy assay (ctDNA sequencing of 105 genes in peripheral blood samples at 5,000x depth) for germline and/or somatic mutations detected in 100 patients (53 AA) receiving androgen deprivation therapy for locally advanced, biochemically recurrent or metastatic PC at Ben Taub Hospital (BTH), a safety net hospital in Harris County/Houston serving a patient population of which 91% are racial/ethnic minorities. For confirmation, we analyzed de-identified NGS data from a nationwide cohort of 1,211 metastatic PC patients (213 AA) previously sequenced with xT and/or xF by Tempus Labs (Chicago, IL). Results: We found higher frequencies of AR (18.9%), TP53 (41.5%), SPOP (20.7%) and homologous recombination repair (HRR) pathway gene mutations, in particular BRCA2 (17%), in our AA BTH cohort, as compared to PC patients of other races/ethnicities. The latter finding was confirmed in the nationwide Tempus Labs cohort, with 91/213 (42.7%) AA patients exhibiting mutation in at least one of 14 HRR pathway genes associated with PC sensitivity to PARP inhibitors, compared to 347/998 (34.7%) non-AA patients (P &lt; 0.05). This difference was mainly driven by higher frequency of BRCA2 (16.9%), CDK12 (8%) and PALB2 (5.2%) mutations in AA patients. In both cohorts, TMPRSS2 fusions were much less common in AA PC patients. Conclusions: The observed high frequency of mutations in key PC drivers in AA patients may reflect differences in disease biology between racial/ethnic groups or the more advanced disease presentation of AA patients due to socioeconomic factors delaying access to healthcare. Our study provides a real-world snapshot of the genomic landscape of advanced PC in a safety net hospital serving large racial/ethnic minority populations and highlights the role that NGS testing can play to improve their access to treatment with novel targeted therapies and to biomarker-based Precision Oncology clinical trials.

Models of Trauma Exposure, Depression, and Suicidality in Safety‐Net Primary Care

Suicidality is a major public health concern, particularly for low-income, trauma-exposed patients with limited access to mental health providers. However, limited research has modeled pathways of suicidality in safety-net primary care samples. Patients (N = 207) in a safety-net primary care clinic completed measures of childhood and adult trauma exposure, depression, and suicidality. Participants (M age = 44.8 years, SD = 11.6), were 60.4% male, 63.8% Black/African American, and predominantly low-income (i.e., 69.1% reported an annual income less than $5,000 USD). Half of the sample reported at least four childhood traumatic events (M = 3.9 events, SD = 3.0) and approximately three adult traumatic events (M = 3.0 events, SD = 2.1). Most participants (82.1%) reported significant depressive symptoms, and 43.5% endorsed recent suicidality. Models showing the mediational effect of depression on the association between trauma exposure and suicidality, β = .20, B = 0.23, SE = 0.05, 95% CI [0.16, 0.32], and the moderational effect of trauma exposure on the association between depression and suicidality, β = .16, B = 0.20, SE = 0.08, p = .007, were both supported. These results underscore the high prevalence of trauma exposure, depression, and suicidality in a safety-net primary care sample. They also highlight the pervasiveness and complexity of suicidality in low-income primary care patients, emphasize the importance of trauma-informed suicide assessment, and identify trauma sequelae and depression as potential treatment targets to reduce suicidality.

Evaluation of the POP-UP programme: a multicomponent model of care for people living with HIV with homelessness or unstable housing.

Homelessness is the greatest risk factor for HIV viremia in San Francisco. Innovative care models for people with HIV (PWH) with homelessness or unstable housing (HUH) are needed to address this inequity. We developed a novel low-barrier clinic-based program for PWH-HUH in an urban safety-net clinic ('POP-UP') and report outcomes on care engagement and viral suppression. A prospective cohort study. San Francisco General Hospital HIV Clinic (Ward 86). We enrolled PWH who are HUH, viraemic and for whom usual care is not working (at least one missed primary care appointment and at least two drop-in visits at Ward 86 in the last year). POP-UP provides drop-in comprehensive primary care, housing assistance and case management, financial incentives and patient navigation with frequent contact. We describe uptake of eligible patients into POP-UP, and cumulative incidence of antiretroviral therapy (ART) initiation, return to care and virologic suppression 6 months post-enrolment, estimated via Kaplan--Meier. Out of 192 referred patients, 152 were eligible, and 75 enrolled. All 75 were off ART and viraemic; 100% had a substance use disorder; and 77% had a mental health diagnosis. Over three-quarters restarted ART within 7 days of enrolment, and 91% returned for follow-up within 90 days. The cumulative incidence of viral suppression at 6 months was 55% (95% confidence interval 43-68). A novel care model for PWH-HUH demonstrates early success in engaging viraemic patients in care and improving viral suppression. Low-barrier, high-contact primary care programmes offering comprehensive services and incentives may improve outcomes for this vulnerable population.

Parental Vaccine Hesitancy and Vaccination Disparities in a Safety-Net System

Poor and minority children still lag behind wealthier, white children in vaccination coverage.1,2 Structural, logistic, cultural, and attitudinal barriers contribute to vaccination disparities.3,4 Vaccine hesitancy (VH) is an important attitudinal barrier and a growing challenge.5 We measured VH and its association with child average days undervaccinated (ADU) in English- and Spanish-speaking parents of children aged 2 years in 3 safety-net clinics, describing associations by language, race, and ethnicity.We surveyed English- and Spanish-speaking parents bringing children aged 2 years to 3 pediatric clinics for well or sick care from August 2019 through February 2020. Parents were eligible if (1) English or Spanish was their first language, (2) their child turned 2 years in the last 12 months, and (3) the child had 1 previous well visit in the last 18 months. We excluded parents of children with medical contraindications to vaccination. A bilingual research assistant recruited dyads in waiting rooms; parents completed the survey on an iPad and received $5. The Colorado Multiple Institutional Review Board approved this study.We measured parental VH with the Parent Attitudes about Childhood Vaccines (PACV) tool because of its previous validation studies,6–8 brevity and simplicity, and rigorous Spanish translation.9 We transformed responses into scaled (0–100) and dichotomized (<50 = not hesitant and ≥50 = hesitant) scores, as described in previous validation work.7 We collected parent demographics and extracted child vaccination data from electronic medical records. Our independent variable of interest was parental VH. Our outcome variable was child ADU at age 2 years (ie, 730 days), which calculates the difference between when vaccine doses are given and when they are due and ranges from 0 (immunized on time) to 638 (unvaccinated).10 We compared parents by VH status on independent variables and performed multivariable linear regression analyses to assess parental VH with child ADU associations. For a 2-sided α of .05, power of 0.80, population ADU mean of 36 ± 89,10 VH effect size of 55 ADU,8,9 and parental VH prevalence of 10%,8,9 we estimated needing 259 parents (24 hesitant and 235 not hesitant). Analyses were conducted by using R (R Foundation for Statistical Computing, Vienna, Austria).Of 293 parents approached, 263 (90%) participated. Two surveys were incomplete, and 6 parents responded twice. In Table 1, we describe our final sample of 255 predominantly Hispanic, English-speaking dyads, stratified by parental VH status. Thirty-three (13%) parents were hesitant; 4 (4%) Spanish-speaking parents were hesitant versus 29 (19%) English-speaking parents (P < .01). PACV item responses also differed by language (Supplemental Table 3). Among all children, the mean ADU at age 2 was 66 days (SD 151). In adjusted analyses, children of English-speaking VH parents had a 114-day greater ADU compared with children of English-speaking non-VH parents (P = .02). Parental VH did not associate with changes in child ADU for Spanish-speaking dyads (P = .61). Children of Black parents (referent to non-Hispanic white) were significantly undervaccinated, independent of parental VH (Table 2).In a safety-net system sample, we found VH in 1 in 8 parents of children aged 2 years, with differential impacts by language, race, and ethnicity. In previous studies, researchers described VH in wealthier, privately insured clinics; in our study, we found VH in clinics serving minority, publicly insured, and Spanish-speaking children, with a prevalence (13%) similar to that previously reported (8%–25%).8,9,11 Nearly 1 in 4 Black parents had VH, and Black children aged 2 years were additionally undervaccinated, independent of parental VH. Parental VH should be explored as an important vaccination barrier along with others identified for Black toddlers, such as public insurance, access, distrust of health care, and missed opportunities.12–15 Parental VH was not associated with ADU changes among Spanish-speaking families, perhaps because of respeto, the cultural value by which Spanish-speaking parents may deferentially subjugate their concerns to clinical expertise.16 This finding bolsters existing work showing how language and culture impact pediatric care.17As a single-site retrospective study of parents of children aged 2 years, our conclusions should be explored in other systems and with other age and language groups. Furthermore, trust between staff and patients may have caused a social desirability bias to participate, and we were unable to recruit dyads who failed to present for care. Still, our work highlights the need to explore VH by language and culture and pursue implementation studies to adapt evidence-based interventions for families in safety-net systems.We acknowledge Jason Glanz, PhD, and the Kaiser Permanente Institute for Health Research for providing us with the algorithm for ADU and instructing us in its use.

Prenatal choline, cannabis, and infection, and their association with offspring development of attention and social problems through 4 years of age.

Prenatal choline is a key nutrient, like folic acid and vitamin D, for fetal brain development and subsequent mental function. We sought to determine whether effects of higher maternal plasma choline concentrations on childhood attention and social problems, found in an initial clinical trial of choline supplementation, are observed in a second cohort. Of 183 mothers enrolled from an urban safety net hospital clinic, 162 complied with gestational assessments and brought their newborns for study at 1 month of age; 83 continued assessments through 4 years of age. Effects of maternal 16 weeks of gestation plasma choline concentrations ⩾7.07 μM, 1 s.d. below the mean level obtained with supplementation in the previous trial, were compared to lower levels. The Attention Problems and Withdrawn Syndrome scales on Child Behavior Checklist 1½-5 were the principal outcomes. Higher maternal plasma choline was associated with lower mean Attention Problems percentiles in children, and for male children, with lower Withdrawn percentiles. Higher plasma choline concentrations also reduced Attention Problems percentiles for children of mothers who used cannabis during gestation as well as children of mothers who had gestational infection. Prenatal choline's positive associations with early childhood behaviors are found in a second, more diverse cohort. Increases in attention problems and social withdrawal in early childhood are associated with later mental illnesses including attention deficit disorder and schizophrenia. Choline concentrations in the pregnant women in this study replicate other research findings suggesting that most pregnant women do not have adequate choline in their diets.

Prevalence and Predictors of Limited Literacy in Public Mental Health Care.

Literacy is an important predictor of health care utilization and outcomes. We examine literacy among people seeking care in a state funded mental health clinic (Site 1) and a safety-net hospital clinic (Site 2). Limited literacy was defined as literacy at or below the 8th grade level. At Site 1, 53% of participants had limited reading literacy and 78% had limited aural literacy. At Site 2, 72% had limited reading and 90% had limited aural literacy. Regression analyses examined associations among limited literacy and psychiatric, neurocognitive and sociodemographic characteristics. Few associations among psychiatric and neurocognitive factors, and literacy were found. At Site 2, black and "other" race participants had higher odds for limited literacy compared to white participants suggesting that limited literacy may be an under-examined mechanism in understanding racial disparities in mental health. Work is needed to understand the relationships among literacy, mental health and mental health care.

Experiences and Perspectives on Adopting New Practices for Social Needs-targeted Care in Safety-net Settings: A Qualitative Case Series Study.

IntroductionClinical settings are being encouraged to identify and address patients’ social needs within the clinic or through partner organizations. The purpose of this qualitative study was to describe the current practice of social needs-targeted care in 3 Texas safety net clinics, and facilitators and barriers to adopting new social needs-targeted care tools and practices.MethodsInterviews were conducted with staff at 3 safety net clinics serving small and mid-sized communities. Analysis focused on perspectives and decisions around adopting new tools or practices related to social needs-targeted care, including standardized screening tools and community resource referral platforms.ResultsNine staff across 3 organizations were interviewed. Two organizations were currently using a standard social needs screening tool in their routine practice, and a third was considering doing so. One organization had adopted a community resource referral platform in partnership with a large community collaboration. Three case studies illustrate a range of facilitators, barriers, perceived benefits, and drawbacks influencing social needs-targeted practices. Benefits of systematic data collection on social needs included the generation of data for community action. Drawbacks include concerns about data privacy. Community resource referral platforms were seen as valuable for creating accountability, but required an influential community partner and adequate community resources. Concerns about disempowering clients and blurring roles were voiced, and potential to increase provider job satisfaction was identified.ConclusionsBenefits and drawbacks of adopting new tools and practices related to social needs-targeted care are strongly influenced by the community context. For the adoption of community resource referral platforms, the outer setting is particularly relevant; adoption readiness is best assessed at the community or regional level rather than the clinic system level. While screening tools are much easier than referral platforms for clinics to adopt, the ability to address identified needs remains heavily based on the outer setting.

Impact of Multi-Component Support Strategies on HIV Virologic Suppression Rates During Covid-19: An Interrupted Time Series Analysis

Background: After COVID-19 shelter-in-place (SIP) orders, viral suppression (VS) rates initially decreased within a safety-net HIV clinic in San Francisco, with greater decreases among people living with HIV (PLWH) experiencing homelessness. We sought to understand if (1) proactive outreach to provide social services, (2) scaling up of in-person visits, and (3) expansion of housing programs could reverse this decline.Methods: We assessed VS 24 months before and 13 months after SIP using mixed-effects logistic regression, followed by interrupted time series (ITS) analysis to examine changes in the rate of viral suppression per month. Loss to follow-up was assessed via active clinic tracing.Findings: Data from 1,816 patients were included in the analysis, with a median age of 51 years, 12% female, 14% experiencing unstable housing/homelessness. After an initial decline with the shelter-in-place order, the adjusted odds of VS increased 1.34-fold over 13 months following institution of the the multi-component strategies (95% confidence interval (CI): 1.21-1.46). In the interrupted time series analysis, the odds of VS continuously increased 1.05-fold per month over the post-intervention period (95% CI: 1.01-1.08). Among PLWH previously experiencing homelessness who were permanently housed or placed in shelter-in-place hotels, the odds of VS were 1.94-fold higher (95% CI:1.05-3.59). The one-year loss to follow-up rate was 2.8 per person-year (95% CI: 2.2-3.5).Interpretation: After an initial destabilization in VS in a large safety-net clinic following SIP orders, the VS rate increased following scale-up of in-person visits, clinic outreach to patients, intensification of social services, and access to COVID-related housing programs during this time. The loss to follow-up rate was similar or lower compared to prior years. Maintaining in-person care for underserved patients, with flexible telemedicine options, along with provision of social services and permanent expansion of housing assistance programs, will be needed to support VS among underserved populations during the COVID-19 pandemic.Funding Information: This work was funded by NIH/NIAID R01AI158013 (M.A.S. M.G. K.A.C, M.O.J), NIH/NIAID R24AI067039 (K.A,C.), and NIH/NIAID P30AI027763 (M.G.)Declaration of Interests: MAS, DVG, MOJ, DH, KAC, SPB and MG report funding from the NIH during conduct of the study. KAC reports investigator-initiated grant support from Gilead Sciences and personal fees from Gilead Sciences outside the submitted work. DVG reports personal fees from Gilead Sciences outside the submitted work. All other authors have nothing to declare. Ethics Approval Statement: This study was approved by the UCSF IRB, and informed consent was not required given processes occurred as part of routine care/quality improvement.

Interest in Clinic-Based Financial Services among Low-Income Prenatal Patients and its Association with Health-Related Social Risk Factors.

BackgroundPoverty and financial stress affect prenatal health and well-being as well as early childhood development. This study sought to examine interest in clinic-based financial services to address financial stress in low-income, Medicaid-enrolled prenatal patients and its relationship with self-reported social risks.MethodsWe conducted a cross-sectional study of patients at a large safety-net prenatal clinic. Participants completed a written survey on interest in linkage to financial services, poverty-related financial stress, difficulty affording social needs, and interest in services to address material hardships. We compared interest in financial and social needs services by level of financial stress using multivariate regression.ResultsRespondents (N = 108) were entirely Medicaid-enrolled, with a majority identifying as Hispanic/Latinx (57%) or Black/African American (20%). Sixty-four percent indicated interest in connection to any of the financial services surveyed. Interest was highest in employment (52%), savings and budgeting (49%), job training/adult education (49%), and financial counseling (48%) services. Individuals with high financial stress, compared to those with low financial stress, expressed a higher level of interest in financial services (aRR = 1.61 [95% CI 1.12-2.39]). Interest in financial services was associated with difficulty affording social needs (aRR = 2.24 [95% CI 1.33-4.43]) and interest in services addressing social needs (aRR = 1.45 [95% CI 1.13-1.92]).ConclusionIn this study of low-income, Medicaid-insured prenatal patients, there was a high degree of interest in clinic-based financial services. Integrating financial services into prenatal health care appears to be an approach that low-income patients would be interested in to directly address poverty and financial stress.

"It was an unexpected bond": How an emerging participant-driven online social network may be enhancing an eLearning nutrition education & supplemental produce intervention.

ObjectiveThe purpose of this paper is to describe a participant-driven, online text message-based social support network that emerged from an eLearning nutrition education and supplemental produce intervention.MethodsAdults (n = 20) who utilized a safety-net clinic for their healthcare participated in a 12-week smartphone-based nutrition education eLearning program using loaned smartphones. Participants also received a box of fresh produce weekly. Participants received weekly text message reminders to collect their produce, and from this researcher-initiated reminder text, a supportive, participant-led, all-group text message thread commenced. Researchers collected all 471 text messages in this all-group thread and included them in the qualitative content analysis of pre and post intervention focus groups.ResultsThe original design of the eLearning nutrition education program was to asynchronously engage learners with nutrition education resources. However, participants themselves initiated a robust group text message support system through which they shared encouragement, recipes, grocery shopping tips, and images of food they prepared with the produce box amongst themselves for the duration of the 12-week intervention.ConclusionThe novel nature by which these participants voluntarily engaged in this peer-to-peer nutrition education-focused text message conversation exemplifies participants becoming agents in their own learning experience and will be used to enhance future eLearning nutrition education experiences developed by our team.

140. Symptoms and Situations Predispose Patients to Use Antibiotics Without Medical Advice

BackgroundUse of antibiotics without a prescription (non-prescription use) contributes to antimicrobial resistance. Non-prescription use includes obtaining and taking antibiotics without a prescription, taking another person’s antibiotics, or taking one’s own stored antibiotics. We conducted a quantitative survey focusing on the factors that impact patients’ decisions to use non-prescription antibiotics.MethodsWe surveyed patients visiting public safety net primary care clinics and private emergency departments in a racially/ethnically diverse urban area. Surveys were read aloud to patients in Spanish and English. Survey domains included patients’ perspectives on which syndromes require antibiotic treatment, their perceptions of health care, and their access to antibiotics without a prescription.ResultsWe interviewed 190 patients, 122 from emergency departments (64%), and 68 from primary care clinics (36%). Overall, 44% reported non-prescription antibiotic use within the past 12 months. Non-prescription use was higher among primary care clinic patients (63%) than the emergency department patients (39%, p = 0.002). The majority felt that antibiotics would be needed for bronchitis (78%) while few felt antibiotics would be needed for diarrhea (30%) (Figure 1).The most common situation identified “in which respondents would consider taking antibiotics without contacting a healthcare provider was “got better by taking this antibiotic before” (Figure 2). Primary care patients were more likely to obtain antibiotics without prescription from another country than emergency department patients (27% vs. 13%, P=0.03). Also, primary care patients were more likely to report obstacles to seeking a doctor’s care, such as the inability to take time off from work or transportation difficulties, but these comparisons were not statistically significant.Figure 1. Patients’ agreement that antibiotics would be needed varied by symptom/syndrome. Figure 2. Situations that lead to non-prescription antibiotic use impacted the two clinical populations differently ConclusionNon-prescription antibiotic use is a widespread problem in the two very different healthcare systems we included in this study, although factors underlying this practice differ by patient population. Better understanding of the factors driving non-prescription antibiotic use is essential to designing patient-focused interventions to decrease this unsafe practice.Disclosures All Authors: No reported disclosures

CSA Partnerships for Health: outcome evaluation results from a subsidized community-supported agriculture program to connect safety-net clinic patients with farms to improve dietary behaviors, food security, and overall health.

Safety-net clinic patients are at risk for food insecurity, which is associated with poor diet quality and negative health outcomes. Research on the impact of interventions addressing food insecurity in health care settings is limited. The primary aim of this study was to determine the preliminary effectiveness of Community-Supported Agriculture (CSA) Partnerships for Health, a 23 week grant-subsidized CSA program, in improving dietary behaviors, self-efficacy to eat vegetables, food security, and overall health among safety-net clinic patients. This single-group pretest-posttest study used interviewer-administered surveys. Study participants were 48 safety-net clinic patients. Outcomes included fruit and vegetable intake, frequency of dark green and orange vegetable intake, self-efficacy to eat vegetables, eating habits, food security status and ability to afford to eat healthy meals, emotional health, social health, and general health status. Differences in values at postintervention versus preintervention were examined using two-sided paired t-tests, McNemar's tests, and generalized estimating equation models. The difference between postintervention versus preintervention values was statistically significant for a number of outcomes, including vegetable intake (p = .030), orange vegetable intake (p = .004), eating habits (p = .039), food security (p = .039), ability to afford to eat healthy meals (p = .003), and general health status (p = .039). Generalized estimating equation models showed similar associations. CSA Partnerships for Health may have the potential to improve dietary behaviors, reduce food insecurity, and improve overall health among safety-net clinic patients.

Does an encounter with the ambulatory clinical pharmacist in the transitional setting play a role in clinic attendance and patient engagement?

BackgroundPatient engagement is an important aspect in improving patient outcomes. However, there is a paucity of literature regarding patient engagement in a nonresearch health care setting and the impacts of ambulatory clinical pharmacist interventions. Patient engagement has several definitions making it challenging to assess, but attendance to initial primary care provider (PCP) visits is an important aspect of patient engagement. ObjectiveThe study objective was to assess if improved patient engagement, defined as attendance to an initial PCP visit, was associated with differences in services provided or pharmacist participation during postdischarge clinic appointments. Practice descriptionThe site of this study was the Johns Hopkins After Care Clinic (JHACC), an interdisciplinary postdischarge, safety net clinic in Baltimore, MD. Practice innovationThrough an interdisciplinary health care team including pharmacists, patients received comprehensive care to assist with postacute disease-state management and transitions of care. Evaluation methodsInitial PCP visit attendance after a postacute visit in a high-risk population was evaluated for differing service delivery factors between groups who attended and did not attend their initial PCP visit. Descriptive statistics and chi-square tests were used for analysis. ResultsPatients were statistically significantly more likely to engage in primary care when clinical pharmacy specialists participated in their JHACC appointment (P = 0.02). Medication education and disease-state counseling improved initial PCP visit attendance, both of which are key pharmacist activities. ConclusionThis study suggests ambulatory clinical pharmacy specialists’ roles in an interdisciplinary clinic model correlates with increased attendance to initial PCP visits, a surrogate for patient engagement. Disease-state education and medication education are both important activities in improving this measure; however, additional research is necessary to determine specific pharmacist interventions associated with patient engagement. As research in patient engagement continues, the positive effects of pharmacist involvement in this area could support their value in ambulatory care services.