Objectives/Goals: The timing of neurosurgery is highly variable for post-hemorrhagic hydrocephalus (PHH) of prematurity. We sought to utilize microvascular imaging (MVI) in ultrasound (US) to identify biomarkers to discern the opportune time for intervention and to analyze the cerebrospinal fluid (CSF) characteristics as they pertain to neurosurgical outcome. Methods/Study Population: The inclusion criteria for the study are admission to the neonatal intensive care unit (NICU) with a diagnosis of Papile grade III or IV. Exclusion criteria are congenital hydrocephalus and hydrocephalus secondary to myelomeningocele/brain tumor/vascular malformation. We are a level IV tertiary referral center. Our current clinical care pathway utilizes brain US at admission and at weekly intervals. Patients who meet certain clinical and radiographic parameters undergo temporary or permanent CSF diversion. Results/Anticipated Results: NEL was implemented at our institution for PHH of prematurity in fall 2022. To date, we have had 20 patients who were diagnosed with grade III or IV IVH, of which 12 qualified for NEL. Our preliminary safety and feasibility results as well as the innovative bedside technique pioneered at our institution are currently in revision stages for publication. Preliminary results of the MVI data have yielded that hyperemia may confer venous congestion in the germinal matrix, which should then alert the neurosurgeon to delay any intervention to avoid progression of intraventricular blood. With regard to CSF characteristics, we anticipate that protein, cell count, hemoglobin, iron, and ferritin will decrease with NEL. Discussion/Significance of Impact: The timing of PHH of prematurity is highly variable. We expect that MVI will offer radiographic biomarkers to guide optimal timing of neurosurgical intervention. A better understanding of CSF characteristics could potentially educate the neurosurgeon with regard to optimal timing of permanent CSF diversion based on specific CSF parameters.

Objectives/Goals: How do we care for a patient whose mental health is deteriorated such that their decision-making capacity may be compromised? The high-potency opioid crisis in Ontario demands that we provide effective care for the affected population. We must also avoid patients having a subjective experience of coercion and must protect their human rights and dignity. Methods/Study Population: Ontario’s legislation governing mental health care will be explored: the Ontario Mental Health, Healthcare Consent, and Substitute Decisions Acts. We will identify best practices/learning across locales. Patients who have been involuntarily treated/confined will be welcomed to voice what the law should contain. International strategies for: coercion reduction practices, advanced care directives, less prohibitionist care culture, and supports for social determinants of health (SDH) may also help Ontarians. Patients, family members, law enforcement, judiciary, community agencies, and healthcare professionals will be invited to contribute via focus groups to drafted mental health care legislative improvements. Thus, we ensure law enforces patient-defined quality care and practical workflows. Results/Anticipated Results: We hope to emerge from our focus group consultation with draft legislative and procedural edits for Ontario’s mental healthcare laws to ensure that the laws protect human rights and that the laws reflect patient-defined needs. We must ensure controls are in place to de-risk power imbalances and limit the incidence of potential procedural misuse. We intend to design legislated procedures to ensure that people don’t get inappropriately involuntarily confined/treated. We will incorporate the perspectives and lived experiences of patients who have experienced involuntary treatment and/or involuntary medical confinement (locally in a focus group(s)) and internationally (in literature) to inform this legislative development. Discussion/Significance of Impact: We will learn from engaged stakeholders about how to shape Ontario’s legislation to support quality mental health care. We hope to identify and draft legislation improvements that voice what patients and their families’ value, drafts informed by evidence-based best practices and informed innovation. Via inclusion, we create a policy that serves.

Objectives/Goals: As translational science (TS) emerges as a field, there is a need for research organizations to understand how to develop capacity for and support the advancement of TS. To support such institutional and infrastructural change, this poster outlines a Translational Science Promotion and Research Capacity (T-SPARC) framework. Methods/Study Population: The T-SPARC framework was developed by members of the Duke University Clinical and Translational Science Institute (CTSI) primarily from CTSI Pilots, Team Science, Evaluation, and Administration, all of whom had identified the need for building institutional capacity for TS at our institution. The group reviewed literature on TS to ensure grounding in current knowledge, drafted an initial TS logic model, and then determined the value of developing a framework addressing building TS institutional capacity. The group then identified other frameworks/models related to behavioral, organizational, and system change; examined scholarship addressing the building of research capacity in colleges and universities; and iterated on a TS-focused framework in multiple working sessions. Results/Anticipated Results: The resultant T-SPARC framework provides a foundation to 1) inform the development of interventions and programs advancing TS and 2) evaluate their effectiveness. It outlines: organizational levels for TS capacity building (large-scale systems, research institutions, teams, and individuals); intervention activities (policies and processes, funding, collaboration and partnership, and training); proximal outcomes (knowledge/attitudes, behaviors, resources/infrastructure, and connections); next-stage outcomes (e.g., interdisciplinary team processes, and research infrastructure); and ultimate goals (fewer translational impediments, improved public health, and health equity). It ingrates TS principles as foundational to, and outcomes of, capacity-building efforts. Discussion/Significance of Impact: T-SPARC, as a framework for building capacity in TS, provides added foundation for advancing the conceptualization and practice of TS. Ultimately, T-SPARC seeks to advance broader goals of reducing longstanding challenges in the translational research process and improving health outcomes.

Objectives/Goals: Our study’s objective is to evaluate RadOnc-GPT, a GPT-4o powered LLM, in generating responses to in-basket messages related to prostate cancer treatment in the Radiation Oncology department. By integrating it with electronic health record (EHR) systems, the goal is to assess its impact on clinician workload, response quality, and efficiency in healthcare communication. Methods/Study Population: RadOnc-GPT was integrated with patient EHRs from both hospital-wide and radiation-oncology-specific databases. The study examined 158 pre-recorded in-basket message interactions from 90 non-metastatic prostate cancer patients. Quantitative natural language processing analysis and two randomized single-blinded grading studies, involving four clinicians and four nurses, were conducted to evaluate RadOnc-GPT’s response quality in completeness, correctness, clarity, empathy, and estimated editing time. Response times were measured to estimate the time saved for clinicians and nurses. The study population included patient messages across all phases of care (pre-, during, and post-treatment) for those undergoing radiotherapy. Results/Anticipated Results: In the single-blinded grader study, clinician graders evaluated 316 responses (158 from human care teams and 158 from RadOnc-GPT). Results showed RadOnc-GPT outperformed human responses in empathy and clarity, while humans excelled in completeness and correctness. Sentiment analyses using TextBlob and VADER revealed RadOnc-GPT responses had a positive mean score of 0.25, whereas human responses clustered around neutral. VADER analysis indicated a high median score for RadOnc-GPT, nearing 1.0, reflecting predominantly positive sentiment, while human responses displayed a broader sentiment range, indicating sensitivity to context. Clinicians averaged 3.60 minutes (SD 1.44) to respond, compared to 6.39 minutes (SD 4.05) for nurses, highlighting RadOnc-GPT’s efficiency in generating timely responses. Discussion/Significance of Impact: RadOnc-GPT effectively generated responses to individualized patient in-basket messages, comparable to those from radiation oncologists and nurses. While human oversight is still necessary to avoid errors, RadOnc-GPT can speed up response times and reduce pressure on care teams, shifting their role from drafting to reviewing responses.

Objectives/Goals: Cerebral amyloid angiopathy-related inflammation (CAA-ri) is a spontaneous inflammatory cerebral vasculopathy that mimics complications of Alzheimer’s disease immunotherapies. Our objective is to evaluate imaging and cerebrospinal fluid (CSF) markers of blood–brain barrier (BBB) impairment and inflammation in CAA-ri. Methods/Study Population: We plan to enroll 20 patients total: 1) 10 patients with CAA-ri as defined by Auriel et al (JAMA Neurology 2016) (exposure group). 2) 10 patients with non-inflammatory CAA defined using Boston criteria 2.0 that do not also meet criteria for CAA-ri (control group). The primary outcome will be Ktrans, a parameter of BBB impairment calculated from dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) of the brain. Secondary outcomes will include DCE-MRI parameter VL, CSF albumin index, CSF fibrinogen, CSF sPDGFR-β, CSF MMP-2, CSF MMP-9, CSF C3, CSF IL1β, CSF IL6, IL8, and TNFα. Statistical comparisons between the exposure and control groups will be made using Wilcoxon rank sum test. Results/Anticipated Results: We anticipate significantly higher levels of BBB impairment and inflammatory biomarkers from DCE-MRI and CSF in subjects with CAA-ri relative to control subjects with non-inflammatory CAA. Discussion/Significance of Impact: Biomarkers are essential to characterize risk factors, pathophysiology, and possible treatment targets in CAA-ri. We plan to use the results of the current study to inform longitudinal studies that will test whether these markers are useful in identifying not only the presence of CAA-ri but also severity, progression, and response to treatment.

Objectives/Goals: Early identification and profiling of planned studies is a critical administrative challenge in providing timely support for clinical trials. Here we describe the collaborative design and development of a clinical trial tracking dashboard to enhance support and quality improvement for investigator-initiated clinical trials at our institution. Methods/Study Population: Trial-CARE organized a workgroup with key stakeholders from WashU business units that manage grants, information technology, bioinformatics, data repository stewardship, and clinical trial support. The workgroup strategized next steps in a “proof-of-concept” effort to determine whether NIH investigator-initiated clinical trial metrics would be accessible via the WashU Data Warehouse. The WashU Data Warehouse is a data repository that pulls in pre-award and post-award data from sources such as the WashU Research-Grant Management System (RMS) and the NIH Reporter Tool. The proof-of-concept findings lead to implementing a plan for Phase 1 of the design, development, and piloting of a visual dashboard to track and offer targeted and timely support to NIH investigator-initiated clinical trials at WashU. Results/Anticipated Results: WashU currently has 217 grants submitted and 114 grants awarded for NIH clinical trials. In the proof of concept, we confirmed and successfully matched over 100 awarded NIH investigator-initiated clinical trials at WashU across RMS and the NIH Reporter Tool. We determined Phase I of the dashboard would track clinical trial data from these two sources via the WashU Data Warehouse. The pilot of Phase I of the dashboard will begin in February 2025. We also identified mission critical data elements not accessible via the WashU Data Warehouse (e.g., enrollment diversity, IND, and IDE). The plan to procure this data will require continued stakeholder support in Phase II of the dashboard to (1) expand data capture in RMS and (2) ingest additional data into the WashU Data Warehouse from RMS or new systems (e.g., ClinicalTrials.gov). Discussion/Significance of Impact: Early access to robust data about NIH investigator-initiated clinical trials via our newly created Phase I dashboard will better position our centralized service cores to support trial success, compliance, and quality improvement across the lifecycle of these clinical trials. In Phase II, we plan to expand the data available in the dashboard.

Objectives/Goals: • To examine current barriers to research translation that collide with funder imperatives to share research results broadly. • To create community-driven guidance on broadening dissemination of research findings to communities to enhance health literacy and trustworthiness in the research process. Methods/Study Population: The Penn State Community Health Equity & Engagement in Research team, part of our Clinical and Translational Science Institute, completed six semi-structured focus groups in the Fall of 2023 (N = 46, including geographically diverse communities with and without research experience). Focus groups included presentation of evidence-based and novel approaches to lay results dissemination, including lay briefs and data walks, to elucidate experiences with and preferences for receiving research study results. Qualitative data were analyzed using MAXQDA software, with successful (>0.70) kappa coefficient achieved for interrater reliability. Codes and themes were developed inductively. Results/Anticipated Results: Focus group characteristics included N = 39 identifying as women, with a mean age of 56 years old and 10% identifying as Black/African American. Geographic breakdown included 49% rural, 44% suburban, and 7% urban. Thematic analyses demonstrated a perceived lack of trustworthiness and representation in the research process, with several cultural and geographic barriers to research accessibility and results dissemination. The majority of participants did not receive research results from prior studies and identified the importance of trusted community messengers to share findings broadly. Participants prefer receiving lay briefs, with health literate infographics, over published study manuscripts. Data walks featuring key findings presented by researchers within communities were highly desired. Discussion/Significance of Impact: Improving translation of study results in underrepresented communities is a catalyst for increasing engagement, demonstrating trustworthiness, and improving health literacy. The development of evidence-based lay dissemination methodologies can increase translation and bolster efforts to support informed, research-ready communities.

Objectives/Goals: The bromodomain PHD finger transcription factor (BPTF) is an oncogenic driver of neuroblastoma. Our objective is to pioneer the discovery of the first class of chemical compounds that engage the PHD finger of BPTF and inhibit its biological function in cellulo, thereby establishing first-in-class chemical probes for this epigenetic reader. Methods/Study Population: Our previous work has identified a collection of small molecules that engage BPTF PHD in vitro. Following structure–activity relationships analysis, candidates will be used in a neuroblastoma cell model to validate BPTF PHD interaction in cellulo and predict therapeutic potential. Nanoluciferase bioluminescence resonance energy transfer (NanoBRET) will be used to confirm BPTF PHD engagement by compounds. Selective toxicity in neuroblastoma cells upon inhibitor treatment will be gauged by comparing cell growth and viability in the IMR-32 cell line against the HEK293 cell line. Treated HEK293 cells will be subjected to the assay for transposase-accessible chromatin (ATAC) and RNA sequencing methods to monitor changes in chromatin structure and transcriptional signatures against untreated cells. Results/Anticipated Results: We hypothesize that compounds with low micromolar potency for BPTF PHD in vitro will engage the target in cellulo and displace NanoLuciferase tagged protein from its HaloTagged® peptide binding partner. Additionally, we anticipate that our inhibitors will show cytotoxicity for IMR-32 cells with limited effects on HEK293 cells. We envision that inhibitor treatment in HEK293 cells will correlate with reduced chromatin exposure, suggesting that blocking the BPTF–histone interaction via PHD finger inhibition hinders the remodeling of transcriptionally silent heterochromatin into a transcriptionally active state. Finally, we expect that inhibitor treatment will result in diminished gene expression of oncogenic transcription factors, including N-Myc, a biomarker of neuroblastoma. Discussion/Significance of Impact: These first-in-class chemical probes for BPTF PHD will enable further investigation of BPTF and high-risk neuroblastoma progression, as well as its role in other diseases. In addition, these compounds will serve as a platform for the development of new anticancer agents that may improve outcomes for children that suffer from high-risk neuroblastoma.

Objectives/Goals: Manual skin assessment in chronic graft-versus-host disease (cGVHD) can be time consuming and inconsistent (>20% affected area) even for experts. Building on previous work we explore methods to use unmarked photos to train artificial intelligence (AI) models, aiming to improve performance by expanding and diversifying the training data without additional burden on experts. Methods/Study Population: Common to many medical imaging projects, we have a small number of expert-marked patient photos (N = 36, n = 360), and many unmarked photos (N = 337, n = 25,842). Dark skin (Fitzpatrick type 4+) is underrepresented in both sets; 11% of patients in the marked set and 9% in the unmarked set. In addition, a set of 20 expert-marked photos from 20 patients were withheld from training to assess model performance, with 20% dark skin type. Our gold standard markings were manual contours around affected skin by a trained expert. Three AI training methods were tested. Our established baseline uses only the small number of marked photos (supervised method). The semi-supervised method uses a mix of marked and unmarked photos with human feedback. The self-supervised method uses only unmarked photos without any human feedback. Results/Anticipated Results: We evaluated performance by comparing predicted skin areas with expert markings. The error was given by the absolute difference between the percentage areas marked by the AI model and expert, where lower is better. Across all test patients, the median error was 19% (interquartile range 6 – 34) for the supervised method and 10% (5 – 23) for the semi-supervised method, which incorporated unmarked photos from 83 patients. On dark skin types, the median error was 36% (18 – 62) for supervised and 28% (14 – 52) for semi-supervised, compared to a median error on light skin of 18% (5 – 26) for supervised and 7% (4 – 17) for semi-supervised. Self-supervised, using all 337 unmarked patients, is expected to further improve performance and consistency due to increased data diversity. Full results will be presented at the meeting. Discussion/Significance of Impact: By automating skin assessment for cGVHD, AI could improve accuracy and consistency compared to manual methods. If translated to clinical use, this would ease clinical burden and scale to large patient cohorts. Future work will focus on ensuring equitable performance across all skin types, providing fair and accurate assessments for every patient.

Objectives/Goals: Endometrial cancer is one of the few cancers that has both a rising incidence and mortality rate. Molecular classification is becoming more important for the management of endometrial cancer but the ability to translate this into clinical practice remains constrained. Our goal is to use AI to predict the molecular subtype from histopathology slides. Methods/Study Population: We utilized the open source endometrial cancer datasets from The Cancer Genome Atlas (TCGA) (N = 387) and Cancer Proteomics Transcriptomic Tumor Analysis Consortium (CPTAC) (N = 135) to develop and train a vision transformer AI model. We used a proprietary cohort of patients (N = 548) for external validation. Whole slide images (WSI) and molecular subtype data were collected. Subtypes include POLE ultramutated (POLE), microsatellite instability (MSI-H), copy-number low (CNV-L), and copy-number high (CNV-H). WSI were preprocessed, and features were extracted. Modified STAMP protocol was used in training, utilizing a pretrained foundation transformer model (Virchow2). Cross-validation of the TCGA was used for initial training, followed by testing on the CPTAC dataset and validation on our proprietary cohort. Results/Anticipated Results: Fivefold cross-validation of the TCGA database (60% training, 20% testing, and 20% validation) developed a best overall model with a mean AUC of 0.74 (POLE 0.78, MSI-H 0.76, CNV-H 0.86, CNV-L 0.77). Overall precision 0.58, recall 0.55. CNV-H was the subtype with the most accurate prediction. CPTAC holdout testing revealed moderately high AUC (POLE 0.63, MSI-H 0.62, CNV-H 0.98, and CNV-L 0.76). Overall precision 0.54 and recall 0.58. Again, CNV-H was the most accurate prediction. Validation on our proprietary cohort revealed a drop in performance with overall mixed results by AUC (POLE 0.50, MSI-H 0.69, CNV-H 0.78, and CNV-L 0.61). Overall precision 0.57, recall 0.45. Again, CNV-H with the most accurate prediction but F1 score dropped from 0.77 in the CPTAC to 0.47 on validation. POLE was the least accurate prediction subtype. Discussion/Significance of Impact: The CNV-H subtype demonstrated robust performance, suggesting the model effectively captures the features associated with this subtype. CNV-L had moderate performance. MSI-H and POLE were notably lower. WSI-based AI models show translational potential for subtype prediction in the management of endometrial cancer but more work is necessary.

Objectives/Goals: The importance of interprofessional communication in healthcare is well documented. However, there has yet to be a theoretical framework for understanding this complex phenomenon. The purpose of this project was to develop a situation-specific theory about the culture of communication in the interprofessional healthcare setting. Methods/Study Population: We conducted individual virtual qualitative interviews with 16 healthcare professionals at different levels and from various disciplines. Participants were practicing clinicians at a tertiary care academic medical center. We transcribed audio-recordings of the interviews verbatim and aligned each transcript for accuracy, confidentiality, and representativeness. Data analysis was conducted using dimensional analysis, including open- and theoretical-coding utilizing an explanatory matrix for theory generation. Results/Anticipated Results: Our findings produced a situation-specific theory on the culture of interprofessional communication in the healthcare setting. Participants described the environments in which they practiced in terms of hierarchical system structures and through a conceptual perspective of resource investment. In other words, they viewed their time and their effort as valuable – and notably, scarce – resources, which they invest in patient care, interprofessional communication, and other tasks. There are two key pathways that characterize the culture of interprofessional communication: collaborative communication and transactional interactions, depending on whether an individual perceives the goals of another person, department, or level of the hierarchical institution as aligned with their own goals. Discussion/Significance of Impact: Positive cultures of interprofessional communication in the healthcare setting depend on perceived goal alignment among individuals, departments, and leadership. Future research can explore how perceptions of goal alignment are developed and empirically test this situation-specific theory in other healthcare system settings.

Objectives/Goals: A key strategy in generating a protective HIV vaccine is the elicitation of broadly neutralizing antibodies (bnAbs), capable of neutralizing a large diversity of HIV-1 isolates. The goal of this study is to identify molecular signatures of HIV bnAb development early in life, to guide the development of a successful pediatric HIV vaccine strategy. Methods/Study Population: We previously defined HIV neutralization breadth in 40 ART-naive children living with HIV. Single-cell RNAseq was performed utilizing peripheral blood mononuclear cells (PBMCs) from the top 5 children with highest neutralization breadth scores and compared their transcriptome to that of PBMCs from 5 children that did not develop neutralization breadth within the first three years of life. Additionally, we incorporated analysis of PBMCs from 5 healthy uninfected children, matched to our experimental groups by race, ethnicity, and gender. Results/Anticipated Results: We expect that a distinct host transcriptional profile is associated with the development of HIV-specific antibody neutralization breadth in early life. Discussion/Significance of Impact: Identifying immune cell transcriptional profiles associated with neutralization breadth will lead to more targeted vaccine approaches for eliciting the appropriate B cell responses and provide an invaluable screening tool allowing early identification of vaccine candidates with the potential to induce bnAbs.

Objectives/Goals: The study is designed to co-develop a genetics education digital platform with community participants. Diverse populations, especially African American communities, are less likely to participate in genetic testing and clinical trials. To increase participation and community input, we want community participants across Louisiana. Methods/Study Population: A cross-sectional, mixed-methods study will be conducted to assess interest in learning about genetics through a digital education platform and to adapt the platform based on participant feedback. Specific Aim 1 will be achieved by recruiting a diverse cohort and collecting demographic data to identify participants’ characteristics. For Specific Aim 2, 35 community participants will be enrolled as a co-design team to complete 5-week educational modules. Feedback from focus groups will guide iterative platform refinements, ensuring the platform is culturally tailored and user-friendly. Results/Anticipated Results: For Specific Aim 1, we anticipate that the majority of participants will express interest in learning about genetics through the digital platform, with demographic data revealing a diverse participant pool, predominantly from African American and Hispanic communities. For Specific Aim 2, we expect that all 35 co-design team members will complete the 5-week modules. Feedback from the focus groups is anticipated to highlight the need for more user-friendly navigation, culturally tailored content, and enhanced visuals. These insights will guide the refinement of the digital platform for improved engagement and accessibility. Discussion/Significance of Impact: This study addresses a critical translational barrier – underrepresentation of African descended communities in genetic testing. By developing a culturally tailored, digital platform to engage these populations, the project aims to reduce health disparities, enhance genetic literacy, and foster inclusivity in genomic research.

Objectives/Goals: We designed a forum to educate participants about bioethical issues in the application of big data (BD) and artificial intelligence (AI) in clinical and translational research (CTR) in underrepresented populations. We sought to determine changes in participants’ interests in ethics, bias, and trustworthiness of AI and BD. Methods/Study Population: 141 individuals registered for the forum, which was advertised to our partner institutions, minority-serving institutions, and community organizations. Registrants received email instructions to complete an AI Trustworthiness (AI-Trust) survey, a questionnaire with integrated qualitative and quantitative measures designed to better understand learners who engaged with the institution-specific AI/Data Science curriculum. Respondents completed the survey using personal devices via a link and QR code, with anonymized responses and enhanced privacy features. 82 people attended; 22 responded to the survey pre-forum and 22 post-forum. Pre- and post-forum responses were qualitatively compared to assess shifts in attitudes toward AI and BD and related interests in ethics, bias, and trustworthiness. Results/Anticipated Results: We found increased interests post- vs. pre-forum in the use of AI for CTR, AI bias and its effects on underrepresented populations, and ethical risk assessment and mitigation strategies for the use of BD to empower research participants. In contrast, trust in AI was lower post- vs. pre-forum. Moreover, respondents also indicated that the current application of AI in healthcare practice would result in increased racial, economic, and gender bias. In comparison, interest in ethical challenges, bioethical considerations, and trustworthiness regarding use of BD and AI in health research and practice did not differ pre- vs. post-forum. Discussion/Significance of Impact: Interest in the application of BD/AI in CTR increased post-forum, but AI distrust and bias expectations also increased, suggesting that learners become more skeptical and discerning as they become more knowledgeable about the complexity of the ethics of AI and BD use in healthcare, especially its application to underrepresented populations.

Objectives/Goals: Youth with IBD have preventive, psychosocial, and acute care needs beyond those of peers, yet receipt of services does not match those needs. Our objectives are to assess the feasibility of (1) an individualized care plan intervention to improve perceived and measured care quality and (2) a pragmatic trial design embedded in pediatric IBD practice. Methods/Study Population: This is a pilot rollout-design randomized trial (n = 60) at a regional academic medical center. Eligible patients are 13–19 years old with IBD for at least 3 months and scheduled for a follow-up visit during the trial. Research staff recruits from one cluster at a time until goal enrollment (14–16). Enrollees are randomized 1:1 to intervention (MyIBD now) or control (MyIBD after the trial). MyIBD combines a tabular summary of individualized acute, chronic, and preventive care needs with nurse facilitator support for patients to use the information. Surveys at baseline, 6 and 12 months measure care quality (Patient Assessment of Chronic Illness Care scale, vaccines, health services) and patient self-management skills (Partners in Health scale). Implementation outcomes are collected via chart review. Results/Anticipated Results: To date, 44 subjects have been randomized. Among subjects, the mean age is 16 years; 73% have Crohn’s disease, 77% have commercial insurance, 75% receive anti-TNF therapy, and 14% live in a rural area. Mean baseline perceived care quality (PACIC scale) is 76.9 (sd 16.3; out of 100); mean baseline perceived self-management skill (PIH scale) is 78.1 (sd 13.4; out of 96). On objective care quality measures, 59% have completed the HPV vaccine series, 32% have received an additional pneumonia vaccine; in the past year 68% have had a screening for mood disorders, 20% an emergency department visit for IBD, and 18% an IBD hospitalization. To date, the IBD clinical team has achieved 100% completion (intervention subjects receive MyIBD plus nurse facilitation) and 0% contamination (control subjects inappropriately receive MyIBD). Discussion/Significance of Impact: Study results to date support the feasibility of the pragmatic, embedded trial design and indicate opportunities for improvement in care quality as perceived by patients and as measured by common preventive and acute care quality indicators. An individualized care plan supported with nurse facilitation may improve pediatric IBD care quality.

Objectives/Goals: Retention in care is vital for people living with HIV. We used human-centered design (HCD) to engage a community-based research panel over a 5-year period, allowing us to incorporate their insights on research guidance and interpretation of findings to investigate correlates of HIV care outcomes. Methods/Study Population: We recruited a diverse panel of individuals who were living with HIV, HIV clinicians, and/or providing non-clinical HIV services in Marion County, Indiana. We conducted biannual sessions using a variety of HCD tools and activities to engage participants. Each session took about three hours, and panelists were compensated for their participation. Due to the COVID-19 pandemic, sessions were initially held virtually. Sessions were designed for project discussion and to facilitate exploration of concerns and challenges facing receipt of HIV services. Our HCD approach put participants in the center of discussion and empowered them to externalize ideas and collaborate meaningfully with our team. Results/Anticipated Results: Since project inception, 48 individuals have joined the panel. Thirty-five are actively engaged, participating in one or more of six sessions conducted to date. We have learned much from the panel. One example is that a residential move might be a risk or protective factor for retention in care and the amount of time one had lived with HIV is a crucial factor. Panel insights have helped guide and prioritize analyses, aided in identification of data missing from our ecosystem, helped interpret results, provided feedback on future interventions, led to a quality improvement project with the local health department, and led to a presentation at a local health equity conference. Discussion/Significance of Impact: Community engagement is essential to impactful and sustainable research. HCD was a successful approach to engage our panel to inform interventions more relevant to the community. We anticipate these methods will be important for others conducting community-engaged research.

Objectives/Goals: This study objective is to evaluate the prevalence and risk factors of burnout in practicing radiologists, with a focus on personal as well as systemic factors. It aims to identify and assess the existing strategies to mitigate burnout, enhance radiologist performance, and improve the quality of patient care. Methods/Study Population: The present study is a systematic review that summarizes existing literature on burnout in radiology, examining its prevalence, risk factors, and effect on diagnostic accuracy, decision-making, and job satisfaction. The review will synthesize validated evidence for emotional exhaustion, depersonalization, and professional fulfillment. The review discusses trends and solutions that have emerged from analysis of data within differing countries, subspecialties, and career stages, focusing on elevated risk of burnout in radiologists. It also assesses downstream effects on patient care quality such as missed diagnoses and increased medical errors. The review also discusses potential strategies for mitigating these negative effects on healthcare delivery. Results/Anticipated Results: The anticipated results of this review are expected to reveal significant variability in burnout rates across radiology subspecialties and practice settings, with prevalence ranging from 33% to 88% (Fawzy et al., 2023). Emotional exhaustion and depersonalization emerge as the most reported symptoms as consistently highlighted in previous studies. Major contributors such as workload, administrative burdens, and technological isolation (e.g., remote work and reduced face-to-face interaction) are anticipated. Radiologists in high-demand areas like interventional radiology and those in private practice may show higher burnout levels than those in academic settings. Protective factors, like exercise, supportive environments, and work-life balance, are expected to reduce burnout levels. Discussion/Significance of Impact: This study calls attention to the importance of addressing radiologist burnout as a key institutional priority. Early and effective interventions are essential for improving job satisfaction, reducing medical errors resulting in enhanced patient care. Addressing burnout is crucial for maintaining a sustainable and effective radiology workflow.

Objectives/Goals: To design and implement programming that better prepares graduate students for diverse roles in a variety of workforce environments, our study models the training landscape and programming needs of graduate students in behavioral, clinical, and biomedical graduate programs at a large Midwestern school of medicine and public health. Methods/Study Population: We conducted six focus groups (two graduate program manager focus groups and four graduate student focus groups), to assess the programming, career development, and training needs of graduate students. Using a grounded theory approach, we first engaged in open coding of a sample of transcripts. After developing a codebook, we continued with an iterative coding process interspersed with coder meetings to discuss emerging and changing codes. Using the framework of landscape analysis allowed our coding and modeling to go beyond graduate student needs and study the varying relationships and contexts that impact graduate students throughout their training, such as relationship to supervisor or institutional policies. Results/Anticipated Results: Preliminary results indicate that students wrestle with their status as both students and workers. Specifically, conflict arises between graduate and supervisor expectations around time spent in class, lab, and other career development activities based on these divergent roles. Students and program managers also note the disparities that arise from the university’s lack of standard, formalized policy on labor issues, such as paid leave. Data also suggest that students on training grants note the difference in access to career development resources compared to colleagues. In many cases, students themselves coordinate ad hoc programming to better suit their career and professional development needs, although this work is not a required aspect of their training. Discussion/Significance of Impact: We characterize current graduate training landscapes, which continue to shift as graduate student bodies diversify, unionize, and express interest in increasingly varied biomedical careers. Data from multiple perspectives facilitate creating, implementing, and evaluating supportive training programs that meet identified student needs.

Objectives/Goals: Predictive performance alone may not determine a model’s clinical utility. Neurobiological changes in obesity alter brain structures, but traditional voxel-based morphometry is limited to group-level analysis. We propose a probabilistic model with uncertainty heatmaps to improve interpretability and personalized prediction. Methods/Study Population: The data for this study are sourced from the Human Connectome Project (HCP), with approval from the Washington University in St. Louis Institutional Review Board. We preprocessed raw T1-weighted structural MRI scans from 525 patients using an automated pipeline. The dataset is divided into training (357 cases), calibration (63 cases), and testing (105 cases). Our probabilistic model is a convolutional neural network (CNN) with dropout regularization. It generates a prediction set containing high-probability correct predictions using conformal prediction techniques, which add an uncertainty layer to the CNN. Additionally, gradient-based localization mapping is employed to identify brain regions associated with low uncertainty cases. Results/Anticipated Results: The performance of the computational conformal model is evaluated using training and testing data with varying dropout rates from 0.1 to 0.5. The best results are achieved with a dropout rate of 0.5, yielding a fivefold cross-validated average precision of 72.19% and an F1-score of 70.66%. Additionally, the model provides probabilistic uncertainty quantification along with gradient-based localization maps that identify key brain regions, including the temporal lobe, putamen, caudate, and occipital lobe, relevant to obesity prediction. Comparisons with standard segmented brain atlases and existing literature highlight that our model’s uncertainty quantification mapping offers complementary evidence linking obesity to structural brain regions. Discussion/Significance of Impact: This research offers two significant advancements. First, it introduces a probabilistic model for predicting obesity from structural magnetic resonance imaging data, focusing on uncertainty quantification for reliable results. Second, it improves interpretability using localization maps to identify key brain regions linked to obesity.

Objectives/Goals: We assessed the feasibility of using a large language model (LLM) to create lay language descriptions of study protocols for recruitment, which has the potential to improve accessibility and transparency of clinical studies and enable participants to make informed decisions. Methods/Study Population: All studies from a clinical research recruitment platform were included, which features human-written lay descriptions and titles for study recruitment. Corresponding protocol summaries in the IRB system were extracted and translated into lay language using a LLM (gpt-35-turbo-0613). A subset was used to develop prompt variations through an iterative process. Prompt strategies evaluated include chain-of-thought and few-shot prompting techniques. LLM-generated and human-written descriptions were compared for readability using Flesch–Kincaid and Simple Measure of Gobbledygook (SMOG) reading grade levels and information completeness using Word Movers’ Distance (WMD). Results/Anticipated Results: A total of 55 study descriptions were included – 10 were used to develop prompts and 45 were used for evaluation. The final LLM instructions included multistep prompts. The LLM was first instructed to produce a two- to three-sentence long description without using scientific jargon and included two pairs of examples. The LLM was then asked to shorten the description and finally to provide an engaging title. LLM-generated and human-written summaries were similar in length (median (IQR) 328 (278.5–360.5) vs. 342 (203–532.5) characters, respectively). LLM-generated summaries had lower Flesch-Kincaid grade level (5.15 vs. 8.28, p Discussion/Significance of Impact: An LLM can be used to generate lay language summaries that are readable at a lower grade level while maintaining semantic similarity. This approach can be used to improve the drafting of summaries for recruitment, thereby improving accessibility to potential participants. Future work includes human evaluation and implementation into practice.