Given substantial reporting delays in overdose deaths, state health departments increasingly use nonfatal overdose data to inform geographically targeted rapid overdose response efforts. We evaluated the extent to which nonfatal overdose events were associated with concurrent and future overdose deaths in Rhode Island. We aggregated nonfatal overdose data from emergency medical services records (2019-2023) and fatal overdose data from the State Unintentional Drug Overdose Reporting System (2020-2023) in 1-, 3-, and 6-month intervals at census block group and census tract levels. Rates of fatal overdose were estimated, relative to nonfatal overdose lagged by 0-12 months, using negative binomial regression, and relative to monthly spikes in nonfatal overdose burden, using zero-inflated Poisson regression. Estimation was implemented using integrated nested Laplace approximation. Each additional nonfatal overdose event per census block group was associated with fatal overdose rates that were 48% higher (95% credible interval, 1.37-1.59) than expected in concurrent months, with smaller associations at the census tract level, in wider time intervals, and when nonfatal overdose data were lagged. Spikes in nonfatal overdose activity were associated with elevated overdose mortality in concurrent periods with fine temporal and geographic granularity, but not in longer time frames and larger geographic areas.

BACKGROUND:HIV continues to disproportionately affect men who have sex with men (MSM) in the United States. Pre-exposure prophylaxis (PrEP) is effective, but disparities persist. Limited studies have conducted systematic evaluations of social determinants of health (SDOH) and their effects on PrEP persistence among MSM. SETTING:We enrolled MSM into a prospective observational cohort to assess progression through the PrEP care continuum. We enrolled patients from 3 diverse settings in the United States from 2018 to 2022. METHODS:We explored the impact of SDOH on PrEP persistence (defined as successfully obtaining PrEP prescriptions and/or clinical documentation of retention in PrEP care) at 6 and 12 months using multilevel, mixed-effects logistic models. RESULTS:A total of N = 300 MSM were enrolled. Median age was 28 years; 40% were Black/African American, and 11% were Hispanic/Latino (H/L). PrEP persistence was 84.7% and 49.3% at 6- and 12-months, respectively. In the unadjusted analysis, Black/African American and H/L individuals were 56% and 54%, respectively, less likely to demonstrate PrEP persistence at 6-and 12-months compared with White/non-H/L individuals. Findings were no longer significant after adjusting for economic stability and educational attainment. Individuals with higher levels of internalized homophobia were less likely to persist on PrEP. Every 1-unit increase on a validated measure of internalized homophobia was independently and negatively associated with PrEP persistence (adjusted odds ratio = 0.95, 95% confidence interval: 0.93 to 0.98). CONCLUSIONS:SDOH are important predictors of racial and ethnic disparities in PrEP persistence among MSM. Addressing these factors could help mitigate racial disparities in PrEP persistence in the United States.

Following federal regulatory changes during the COVID-19 pandemic, Rhode Island expanded methadone access for opioid treatment programs (OTPs) in March 2020. The policy, which permitted take-home dosing for patients, contrasted with longstanding restrictions on methadone. This study used patient-level OTP admission and discharge records to compare six-month retention before and after the policy change. We conducted a retrospective cohort study of 1,248 patients newly admitted to OTPs between March 18 and June 30 of 2019 (pre-policy) and 2020 (post-policy). We used logistic regression to estimate associations with retention before and after the policy and used a machine learning approach, the Heterogeneous Treatment Effect (HTE)-Scan, to explore heterogeneity in retention across subgroups. Overall, we found no change in retention following the policy, with an adjusted OR of 1.08 (95% CI: 0.80-1.45) and adjusted RR of 1.03 (0.90-1.18). Using HTE-Scan, we identified two subgroups with significantly increased retention above the overall cohort: (1) patients with below high school education and past-month arrest and (2) male, non-Hispanic white or Hispanic/Latino patients reporting heroin or fentanyl use with past-month arrest. We identified no subgroups with significantly decreased retention. Collectively, findings suggest that expanded methadone access may benefit vulnerable populations without harming overall retention.

Background Neighborhood-level overdose risk may vary over time. In Rhode Island, we developed and validated a machine learning model to identify the 20 percent of census block groups (CBGs) at the highest predicted risk of future overdose death. We updated this model periodically between November 2021 and August 2024 to generate six sets of predictions. This study aims to characterize the trajectory of each CBG’s predicted overdose risk over time across these six periods. Methods In each prediction period, CBGs were designated as “high risk” or not designated as “high risk” based on our model’s 20 percent predicted overdose risk threshold. We implemented sequence analysis to describe unique trajectories in each CBG’s risk designation over each prediction period. We then calculated optimal matching distances to estimate dissimilarity between each pair of trajectories and applied agglomerative hierarchical clustering to group similar trajectories. Results The 809 CBGs included in this study followed 60 unique trajectories in predicted overdose risk designation over the six prediction periods. Clustering of trajectories favored a solution with five trajectory groups. Most CBGs (73.4 %) were rarely or never designated as “high risk”, 7.9 % of CBGs were always designated as “high risk”, and the remaining 18.7 % were designated as “high risk” in multiple prediction periods, represented by trajectory groups with different patterning over time. Conclusions Given the substantial variability in which CBGs were at highest overdose risk over time, dynamic machine learning predictions may inform harm reduction resource allocation by identifying neighborhoods with emerging needs.

To meet the needs of diverse communities, public health authorities are increasingly reliant on hyperlocal interventions targeting specific health issues and distinct populations. To facilitate epidemiological evaluation of hyperlocal interventions on community-level outcomes, we developed a framework of six practice-based considerations for researchers: spatial zone of impact, temporal resolution of impact, outcome of interest, definition of a plausible comparison group, micro vs. macro impacts, and practitioner engagement. We applied this framework to a case study of an impact evaluation of the New York City (NYC) overdose prevention centers (OPCs) on neighborhood-level drug-related arrests. We used drug arrest data from NYC from January 1, 2014, to September 30, 2023 and US Census data to conduct synthetic control modeling, comparing pre- and post-OPC arrests in the neighborhoods surrounding the two NYC OPCs (East Harlem and Washington Heights). We conducted sensitivity analyses to validate our results and compare our findings with those from a prior published study. Our findings indicate no significant change in drug-related arrests following the OPC openings. The mean absolute differences in daily drug-related arrests between the OPCs and their synthetic controls were 0.63 (p = 0.19) in East Harlem and 0.14 (p = 0.22) in Washington Heights. Sensitivity analyses corroborated our main results. Overall, findings demonstrate how our framework can be used to guide future epidemiological evaluations of diverse, hyperlocal public health interventions.

OBJECTIVES:In Rhode Island, drug overdose deaths increased by 28% in the first 6 months of the COVID-19 pandemic in 2020 as compared with the previous year (2019), mirroring national trends. We explored how the spatial distribution of overdose deaths overlapped with that of COVID-19 cases to identify levels and increased prevalence of these health issues among census tracts in Rhode Island. METHODS:We used data from the Rhode Island Department of Health and the US Census Bureau to calculate annualized COVID-19 case rates (from March 20, 2020, through December 31, 2021) and unintentional overdose death rates by census tract (from January 1, 2018, through December 31, 2021). We used bivariate cluster analyses to group census tracts into clusters of high-high, low-low, high-low, and low-high overdose deaths and COVID-19 case rates per 100 000 population. RESULTS:Clusters with high overdose death rates and high COVID-19 case rates were identified in urban census tracts around the capital city of Providence, whereas clusters with low overdose death rates and low COVID-19 case rates were identified in the state's southern census tracts. Structural factors differed among cluster groups: cluster groups with high overdose death rates and high COVID-19 case rates had greater percentages of households with overcrowding (mean [SD] = 1.6% [1.0%]), people living below the federal poverty level (17.5% [7.4%]), and people with a high school degree or less (37.8% [7.8%]) than the other cluster groups. CONCLUSIONS:Targeted investments in community-led and place-based public health interventions can be used to address underlying social and structural determinants of health (eg, overcrowding, poverty, low education levels) in communities with high rates of overdose deaths and COVID-19 cases.

Context: Predictive modeling can identify neighborhoods at elevated risk of future overdose death and may assist community organizations’ decisions about harm reduction resource allocation. In Rhode Island, PROVIDENT is a research initiative and randomized community intervention trial that developed and validated a machine learning model that predicts future overdose at a census block group (CBG) level. The PROVIDENT model prioritizes the top 20th percentile of CBGs at highest risk of future overdose death over the subsequent 6-month period. In CBGs assigned to the trial intervention arm, these predictions are then displayed for partnering community organizations via an interactive mapping dashboard. Objective: To evaluate whether CBGs prioritized by the PROVIDENT model were associated with increased user engagement via an online dashboard for fatal overdose forecasting and resource planning. Design: We estimated prevalence ratios using modified Poisson regression models, adjusted for CBG-level characteristics that may confound the relationship between model predictions and dashboard engagement. Setting: We used CBG-level data in Rhode Island (N = 809) from November 2021 to July 2024. Intervention: Our exposure of interest was whether each CBG was prioritized by the PROVIDENT model and shown as prioritized on the interactive mapping dashboard. Main Outcome Measure: Our primary outcome was whether a dashboard user from any partnering community organization engaged (eg, clicked, interacted with dashboard elements, or completed assessment or planning surveys) with each CBG on the interactive mapping dashboard. Results: After adjusting for previous model predictions and dashboard engagement, nonfatal overdose counts, and distribution of race and ethnicity, poverty, unemployment, and rent burden, dashboard users were 1.0 to 2.4 times as likely to engage with CBGs prioritized by the PROVIDENT model that were shown as prioritized on the dashboard as compared to CBGs that were prioritized by the PROVIDENT model that were blinded on the dashboard. Conclusions: Interactive mapping tools with predictive modeling may be useful to support community-based harm reduction organizations in the allocation of resources to neighborhoods predicted to be at high risk of future overdose death.

Oral HIV pre-exposure prophylaxis (PrEP) is highly effective for preventing HIV. Several different developments in the US either threaten to increase or promise to decrease PrEP out-of-pocket costs and access in the coming years. In a sample of 58,529 people with a new insurer-approved PrEP prescription, we estimated risk-adjusted percentages of patients who abandoned (did not fill) their initial prescription across six out-of-pocket cost categories. We then simulated the percentage of patients who would abandon PrEP under hypothetical changes to out-of-pocket costs, ranging from $0 to more than $500. PrEP abandonment rates of 5.5 percent at $0 rose to 42.6 percent at more than $500; even a small increase from $0 to $10 doubled the rate of abandonment. Conversely, abandonment rates that were 48.0 percent with out-of-pocket costs of more than $500 dropped to 7.3 percent when those costs were cut to $0. HIV diagnoses were two to three times higher among patients who abandoned PrEP prescriptions than among those who filled them. These results imply that recent legal challenges to the provision of PrEP with no cost sharing could substantially increase PrEP abandonment and HIV rates, upending progress on the HIV/AIDS epidemic.

Uptake and retention in clinical care for pre-exposure prophylaxis (PrEP) is suboptimal, particularly among young African American men who have sex with men (MSM) in the Deep South. We conducted a two-phase study to develop and implement an intervention to increase PrEP persistence. In Phase I, we conducted focus groups with 27 young African American MSM taking PrEP at a community health center in Jackson, Mississippi to elicit recommendations for the PrEP persistence intervention. We developed an intervention based on recommendations in Phase I, and in Phase II, ten participants were enrolled in an open pilot. Eight participants completed Phase II study activities, including a single intervention session, phone call check-ins, and four assessments (Months 0, 1, 3, and 6). Exit interviews demonstrated a high level of acceptability and satisfaction with the intervention. These formative data demonstrate the initial promise of a novel intervention to improve PrEP persistence among young African American MSM.

Black sexually minoritized men (SMM) and transgender women (TW) are subgroups with lower rates of substance use and comparable rates of condom use relative to White SMM and TW yet experience heightened vulnerability to HIV. This study sought to explore associations of substance use, including sex-drug use (i.e., drug or alcohol use during sex to enhance sex), and condomless sex among Black SMM and TW. Data were collected from Black SMM and TW living in Chicago, Illinois, enrolled in the Neighborhoods and Networks (N2) cohort study, from November 2018 to April 2019. We used bivariate analyses followed by a multilevel egocentric network analysis to identify factors associated with condomless sex. We conducted Spearman correlation coefficients to examine correlations between pairs of sex-drugs to enhance sex. We used a bipartite network analysis to identify correlates of sex-drug use and condomless sex. A total of 352 Black SMM and TW (egos) provided information about 933 sexual partners (alters). Of respondents, 45

OBJECTIVE:Examine differences in neighborhood characteristics and services between overdose hotspot and non-hotspot neighborhoods and identify neighborhood-level population factors associated with increased overdose incidence. METHODS:We conducted a population-based retrospective analysis of Rhode Island, USA residents who had a fatal or non-fatal overdose from 2016 to 2020 using an environmental scan and data from Rhode Island emergency medical services, State Unintentional Drug Overdose Reporting System, and the American Community Survey. We conducted a spatial scan via SaTScan to identify non-fatal and fatal overdose hotspots and compared the characteristics of hotspot and non-hotspot neighborhoods. We identified associations between census block group-level characteristics using a Besag-York-Mollié model specification with a conditional autoregressive spatial random effect. RESULTS:We identified 7 non-fatal and 3 fatal overdose hotspots in Rhode Island during the study period. Hotspot neighborhoods had higher proportions of Black and Latino/a residents, renter-occupied housing, vacant housing, unemployment, and cost-burdened households. A higher proportion of hotspot neighborhoods had a religious organization, a health center, or a police station. Non-fatal overdose risk increased in a dose responsive manner with increasing proportions of residents living in poverty. There was increased relative risk of non-fatal and fatal overdoses in neighborhoods with crowded housing above the mean (RR 1.19 [95 % CI 1.05, 1.34]; RR 1.21 [95 % CI 1.18, 1.38], respectively). CONCLUSION:Neighborhoods with increased prevalence of housing instability and poverty are at highest risk of overdose. The high availability of social services in overdose hotspots presents an opportunity to work with established organizations to prevent overdose deaths.

Background: Drug overdose persists as a leading cause of death in the United States, but resources to address it remain limited. As a result, health authorities must consider where to allocate scarce resources within their jurisdictions. Machine learning offers a strategy to identify areas with increased future overdose risk to proactively allocate overdose prevention resources. This modeling study is embedded in a randomized trial to measure the effect of proactive resource allocation on statewide overdose rates in Rhode Island (RI). Methods: We used statewide data from RI from 2016 to 2020 to develop an ensemble machine learning model predicting neighborhood-level fatal overdose risk. Our ensemble model integrated gradient boosting machine and super learner base models in a moving window framework to make predictions in 6-month intervals. Our performance target, developed a priori with the RI Department of Health, was to identify the 20% of RI neighborhoods containing at least 40% of statewide overdose deaths, including at least one neighborhood per municipality. The model was validated after trial launch. Results: Our model selected priority neighborhoods capturing 40.2% of statewide overdose deaths during the test periods and 44.1% of statewide overdose deaths during validation periods. Our ensemble outperformed the base models during the test periods and performed comparably to the best-performing base model during the validation periods. Conclusions: We demonstrated the capacity for machine learning models to predict neighborhood-level fatal overdose risk to a degree of accuracy suitable for practitioners. Jurisdictions may consider predictive modeling as a tool to guide allocation of scarce resources.

CONTEXT:Data dashboards have emerged as critical tools for surveillance and informing resource allocation. Despite their utility and popularity during COVID-19, there is a growing need to understand what tools and training are tailored to nonprofit community-based organizations that may partner with public health officials. PROGRAM:In June 2021, the Rhode Island Department of Health and Brown University partnered to create Project SIGNAL (Spatiotemporal Insights to Guide Nuanced Actions Locally), which utilizes spatiotemporal analytics to identify Rhode Island's largest disparities in COVID-19-related outcomes (eg, testing, diagnosis, vaccinations) at the neighborhood level. Results were hosted in an interactive online dashboard (signal-ri.org) designed using principles of the CDC Clear Communication Index. The target audience included a network of 15 geographic areas called Health Equity Zones, funded by the health department to provide critical grassroots public health programs to address social, health, and economic outcomes in their communities. IMPLEMENTATION:To disseminate the dashboard, a 6-hour virtual workshop series was created to train leaders to use the dashboard and increase their confidence in understanding common public health data terminology and concepts and better prepare attendees for rapid decision making during future public health emergencies. EVALUATION:The Project SIGNAL dashboard was launched in August 2022 and has been accessed over 7500 times. A total of 84 community leaders were trained to use this dashboard, increasing their confidence in applying common public health metrics to make decisions about their COVID-19-related activities. DISCUSSION:While several studies have outlined best practices for data dashboards, this is among the first to examine incorporating these practices into a spatiotemporal decision tool designed specifically for community organizations. Project SIGNAL demonstrates that by incorporating design best practices and pairing data dashboards with hands-on training, we can empower community leaders to utilize advanced spatiotemporal methods to identify health disparities and take localized action.

Oral HIV pre -exposure prophylaxis (PrEP) is highly effective for preventing HIV. Several different developments in the US either threaten to increase or promise to decrease PrEP out-of-pocket costs and access in the coming years. In a sample of 58,529 people with a new insurerapproved PrEP prescription, we estimated risk -adjusted percentages of patients who abandoned (did not fill) their initial prescription across six out-of-pocket cost categories. We then simulated the percentage of patients who would abandon PrEP under hypothetical changes to out-ofpocket costs, ranging from $0 to more than $500. PrEP abandonment rates of 5.5 percent at $0 rose to 42.6 percent at more than $500; even a small increase from $0 to $10 doubled the rate of abandonment. Conversely, abandonment rates that were 48.0 percent with out-of-pocket costs of more than $500 dropped to 7.3 percent when those costs were cut to $0. HIV diagnoses were two to three times higher among patients who abandoned PrEP prescriptions than among those who filled them. These results imply that recent legal challenges to the provision of PrEP with no cost sharing could substantially increase PrEP abandonment and HIV rates, upending progress on the HIV/AIDS epidemic.

Importance Clinicians are a key component of preexposure prophylaxis (PrEP) care. Yet, no prior studies have quantitatively investigated how PrEP adherence differs by clinician specialty. Objective To understand the association between prescribing clinician specialty and patients not picking up (reversal/abandonment) their initial PrEP prescription. Design, Setting, and Participants This cross-sectional study of patients who were 18 years or older used pharmacy claims data from 2015 to 2019 on new insurer-approved PrEP prescriptions that were matched with clinician data from the US National Plan and Provider Enumeration System. Data were analyzed from January to May 2022. Main Outcomes and Measures Clinician specialties included primary care practitioners (PCPs), infectious disease (ID), or other specialties. Reversal was defined as a patient not picking up their insurer-approved initial PrEP prescription. Abandonment was defined as a patient who reversed and still did not pick their prescription within 365 days. Results Of the 37 003 patients, 4439 (12%) were female and 32 564 (88%) were male, and 77% were aged 25 to 54 years. A total of 24 604 (67%) received prescriptions from PCPs, 3571 (10%) from ID specialists, and 8828 (24%) from other specialty clinicians. The prevalence of reversals for patients of PCPs, ID specialists, and other specialty clinicians was 18%, 18%, and 25%, respectively, and for abandonments was 12%, 12%, and 20%, respectively. After adjusting for confounding, logistic regression models showed that, compared with patients who were prescribed PrEP by a PCP, patients prescribed PrEP by ID specialists had 10% lower odds of reversals (odds ratio [OR], 0.90; 95% CI, 0.81-0.99) and 12% lower odds of abandonment (OR, 0.88; 95% CI, 0.78-0.98), while patients prescribed by other clinicians had 33% higher odds of reversals (OR, 1.33; 95% CI, 1.25-1.41) and 54% higher odds of abandonment (OR, 1.54; 95% CI, 1.44-1.65). Conclusion The results of this cross-sectional study suggest that PCPs do most of the new PrEP prescribing and are a critical entry point for patients. PrEP adherence differs by clinician specialties, likely due to the populations served by them. Future studies to test interventions that provide adherence support and education are needed.

Background: Acute reperfusion treatment reduces morbidity in Acute Ischemic Stroke (AIS). National guidelines use Last Known Well (LKW) estimates to determine eligibility. When LKW and symptom onset (SO) are “coupled,” treatment eligibility determination is easier. However, eligibility becomes complex when times are “decoupled,” such as in wake-up strokes. We aimed to understand social differences in AIS patients with coupled vs decoupled LKW and SO times. Methods: Time is Brain is an ongoing, multicenter study assessing effects of social network dynamics on pre-hospital delay in socioeconomically disadvantaged stroke patients. Demographics, stroke characteristics, tPA administration, LKW, and SO times were collected from electronic medical records or interviews from Jan to Jul 2023. Symptom onset-to-door (OTD) times were calculated. Fourteen out of seventy patients diagnosed with intracerebral hemorrhage or transient ischemic attack were excluded. Fifty-six were divided into Coupled or Decoupled LKW/SO conditions. Mann-Whitney U , and Fisher’s exact or Pearson X 2 tests were performed. Significance was defined as p<0.05. Results: Fifty-six patients diagnosed with AIS have been enrolled (mean age 67.0± 14.3, 54.3% Female, median NIHSS 2). 87.5% (n=49) self-identified as non-Hispanic and 17.9% (n=10) as Black. Fourteen patients had coupled and forty-two had decoupled LKW/SO times. Among those with decoupled times, 81.0% (n=34) had onset at home and 26.2% (n=11) awoke with symptoms. Cortical AIS was prevalent in both groups (58.8% v. 41.5%, p-value: 0.815). EMS use (58.8% v. 60.4%, p-value: 0.762) and tPA administration (23.5% v. 18.9%, p-value: 1.000) were comparable. Decoupled LKW/SO patients less commonly presented with dysarthria than coupled patients (29.4% v. 54.7%, p=0.036). Median OTD time of coupled v. decoupled groups was 458.5 v.105 min (IQR: 101.25-996.75 v. 60-390.0; p-value: 0.048). Conclusions: In our preliminary analysis of this multicenter cohort, three out of every four AIS stroke patients presented with decoupled LKW/SO times. Decoupling commonly occurred outside the wake-up stroke setting. Ongoing studies will discern patient-level social and clinical characteristics that describe these two important AIS populations.

As the field of public health rises to the demands of real-time surveillance and rapid data-sharing needs in a postpandemic world, it is time to examine our approaches to the dissemination and accessibility of such data. Distinct challenges exist when working to develop a shared public health language and narratives based on data. It requires that we assess our understanding of public health data literacy, revisit our approach to communication and engagement, and continuously evaluate our impact and relevance. Key stakeholders and cocreators are critical to this process and include people with lived experience, community organizations, governmental partners, and research institutions. In this viewpoint paper, we offer an instructive approach to the tools we used, assessed, and adapted across 3 unique overdose data dashboard projects in Rhode Island, United States. We are calling this model the “Rhode Island Approach to Public Health Data Literacy, Partnerships, and Action.” This approach reflects the iterative lessons learned about the improvement of data dashboards through collaboration and strong partnerships across community members, state agencies, and an academic research team. We will highlight key tools and approaches that are accessible and engaging and allow developers and stakeholders to self-assess their goals for their data dashboards and evaluate engagement with these tools by their desired audiences and users.

BACKGROUND:Cancer screening is effective in reducing the burden of breast, cervical, and colorectal cancers, but not all communities have appropriate access to these services. In this study, we aimed to identify under-resourced communities by assessing the association between the Social Vulnerability Index (SVI) with screening rates for breast, cervical, and colorectal cancers in ZIP-code tabulation areas (ZCTAs) in Rhode Island. METHODS:This study leveraged deidentified health insurance claims data from HealthFacts RI, the state's all-payer claims database, to calculate screening rates for breast, cervical, and colorectal cancers using Healthcare Effectiveness Data and Information Set measures. We used spatial autoregressive Tobit models to assess the association between the SVI, its four domains, and its 15 component variables with screening rates in 2019, accounting for spatial dependencies. RESULTS:In 2019, 73.2, 65.0, and 66.1% of eligible individuals were screened for breast, cervical, and colorectal cancer, respectively. For every 1-unit increase in the SVI, screening rates for breast and colorectal cancer were lower by 0.07% (95% CI 0.01-0.08%) and 0.08% (95% CI 0.02-0.15%), respectively. With higher scores on the SVI's socioeconomic domain, screening rates for all three types of cancers were lower. CONCLUSION:The SVI, especially its socioeconomic domain, is a useful tool for identifying areas that are under-served by current efforts to expand access to screening for breast, cervical, and colorectal cancer. These areas should be prioritized for new place-based partnerships that address barriers to screening at the individual and community level.

Identifying county-level factors that influence pre-exposure prophylaxis (PrEP) adherence is critical for ending the HIV epidemic in the United States (US). PrEP primary reversal is a term used to describe patients who do not obtain their prescribed medication from the pharmacy. This study sought to identify factors associated with PrEP reversal at the county level in 2018. Data were collected from Symphony Health Analytics, AIDS Vu, the US Census Bureau, and the Centers for Disease Control and Prevention National Prevention Information Network. Bivariate Choropleth maps were created to identify counties with high and low levels of PrEP reversal and HIV incidence. This was followed by bivariate analysis to determine the association between predictor variables and percent PrEP reversal. Finally multivariable logistic regressions were used to assess the association between percent PrEP reversal and variables that were significant from the bivariate analysis. A total of 308 counties were included in this analysis, where the mean number of PrEP prescriptions for counties was 44, with a median of 14 (Interquartile range 7-34). In the multivariable analysis, counties with higher level of unemployment (aOR: 1.10, 95% CI: 1.05-1.16) and rural counties (1.10: 1.04-1.17) had higher odds of PrEP reversal; while counties with higher household crowding (0.97: 0.95-0.99) had lower odds of PrEP reversal. Findings show the need for expanding and implementing programs as well as policies to improve PrEP services that are tailored to local socioeconomic circumstances.

Introduction: 80% of strokes occur in the company of others and 70% take place at home. However, research on prehospital delay has overlooked the social nature of stroke detection and triage. In the Time is Brain study, we aim to understand how a patient’s social network affects the decision to seek medical care during stroke symptom onset. Methods: We conducted in-depth interview in 78 patients (ages 30 to 90 years) in the longitudinal, multicenter Time is Brain study. During the interview, we mapped and examined relationships and interactions between patients and individuals involved in the decision-making process during the stroke. We enrolled during their initial stroke hospitalization at Mass General Brigham or Yale New Haven Hospital. Eligibility criteria include having an acute stroke syndrome presentation that may include diagnoses of acute ischemic stroke, intracerebral hemorrhage (ICH), or transient ischemic attack (TIA). Trauma-related syndromes, such as subdural hemorrhage, were excluded. Results: 82.5% (n=66) of patients were accompanied by at least one other person during symptom onset. 60% (n=48) contacted at least one other person. 82.5% (n=66) of strokes occurred in a home-setting and 17.5% (n=14) occurred at work or another public space. 36.25% (n=29) of patients identified their spouse as the first person made aware of symptoms. Men were more likely than women to be accompanied by a spouse (X 2 = 6.8211, p-value = 0.03302). Conclusions: Building on prior studies that have highlighted the important role of the social environment in stroke detection and triage, this is a high-resolution study of the social agents and behaviors driving stroke decision-making. These findings highlight the active and often lengthy communication, and unspoken observations that shape layperson decisions during stroke crises. These data are important to identify mechanisms of prehospital delay and build social network-based interventions to improve time-to-treatment.