562 Mapping and navigating translational resources with generative AI
Objectives/Goals: Translational researchers often struggle to navigate a complex constellation of institutional resources spanning the IRB to bioinformatics units. We had two aims 1) Systematically map all institution-wide research support units and 2) leverage this database within a generative AI v...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Cambridge University Press
2025-04-01
|
| Series: | Journal of Clinical and Translational Science |
| Online Access: | https://www.cambridge.org/core/product/identifier/S2059866124011336/type/journal_article |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849393789544693760 |
|---|---|
| author | Jonathan Gelfond Meredith Zozus Martin Goros Jennifer Potter Kimberly K. Summers Susanne Schmidt Stephanie Rowan Laura Aubree Shay |
| author_facet | Jonathan Gelfond Meredith Zozus Martin Goros Jennifer Potter Kimberly K. Summers Susanne Schmidt Stephanie Rowan Laura Aubree Shay |
| author_sort | Jonathan Gelfond |
| collection | DOAJ |
| description | Objectives/Goals: Translational researchers often struggle to navigate a complex constellation of institutional resources spanning the IRB to bioinformatics units. We had two aims 1) Systematically map all institution-wide research support units and 2) leverage this database within a generative AI virtual concierge tailored to local investigator queries and needs. Methods/Study Population: This study leveraged mixed methods approach. First, we conducted needs assessments of local study teams to identify barriers to translation, revealing that research resources are often unknown to study teams. Second, we identified all investigators, institutional units, and offices offering such resources that we call research support units (RSUs). RSUs were surveyed, collecting contact information (leadership, website, physical location), services provided, type of research supported, and performance metrics. Third, the resource database was integrated into a large language model (LLM, e.g., ChatGPT4o) using a retrieval augmented generation (RAG) system within an R Shiny application called virtual concierge. Queries and responses are recorded for quality improvement. Results/Anticipated Results: Needs assessment focus groups consisted of clinical and basic science investigators, study team members (e.g., clinical research assistants), core directors, and administrators (n = 26). Six sessions were conducted in Spring 2024. A major resultant theme was difficulty finding RSUs “by trial and error” and lacking a “clear defined pathway” for accessing RSUs. This prompted a survey-based environmental scan to identify institutional research resources. There were 122 diverse RSUs ranging from the IRB, to grant writing, to single cell sequencing. Each research unit offered a median of 6 service types, totaling 410 service types overall. The resultant Virtual Concierge meaningfully responds to investigator resource queries with appropriate contact and access information. Usability testing is underway. Discussion/Significance of Impact: Linking researchers with translational resources requires mutual understanding, timely communication, and coordination across teams. We systematically filled these information gaps between investigators and institutional resources. Our Virtual Concierge AI bot can help researchers navigate resources through the translational process. |
| format | Article |
| id | doaj-art-1532a80646d64463b055e36041ca0c3e |
| institution | Kabale University |
| issn | 2059-8661 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | Cambridge University Press |
| record_format | Article |
| series | Journal of Clinical and Translational Science |
| spelling | doaj-art-1532a80646d64463b055e36041ca0c3e2025-08-20T03:40:18ZengCambridge University PressJournal of Clinical and Translational Science2059-86612025-04-01916516510.1017/cts.2024.1133562 Mapping and navigating translational resources with generative AIJonathan Gelfond0Meredith Zozus1Martin Goros2Jennifer Potter3Kimberly K. Summers4Susanne Schmidt5Stephanie Rowan6Laura Aubree Shay7UT Health San AntonioUT Health San AntonioUT Health San AntonioUT Health San AntonioUT Health San AntonioUT Health San AntonioUT Health San AntonioUT Health Houston School of Public Health Muayad Hamidi, UT Health San AntonioObjectives/Goals: Translational researchers often struggle to navigate a complex constellation of institutional resources spanning the IRB to bioinformatics units. We had two aims 1) Systematically map all institution-wide research support units and 2) leverage this database within a generative AI virtual concierge tailored to local investigator queries and needs. Methods/Study Population: This study leveraged mixed methods approach. First, we conducted needs assessments of local study teams to identify barriers to translation, revealing that research resources are often unknown to study teams. Second, we identified all investigators, institutional units, and offices offering such resources that we call research support units (RSUs). RSUs were surveyed, collecting contact information (leadership, website, physical location), services provided, type of research supported, and performance metrics. Third, the resource database was integrated into a large language model (LLM, e.g., ChatGPT4o) using a retrieval augmented generation (RAG) system within an R Shiny application called virtual concierge. Queries and responses are recorded for quality improvement. Results/Anticipated Results: Needs assessment focus groups consisted of clinical and basic science investigators, study team members (e.g., clinical research assistants), core directors, and administrators (n = 26). Six sessions were conducted in Spring 2024. A major resultant theme was difficulty finding RSUs “by trial and error” and lacking a “clear defined pathway” for accessing RSUs. This prompted a survey-based environmental scan to identify institutional research resources. There were 122 diverse RSUs ranging from the IRB, to grant writing, to single cell sequencing. Each research unit offered a median of 6 service types, totaling 410 service types overall. The resultant Virtual Concierge meaningfully responds to investigator resource queries with appropriate contact and access information. Usability testing is underway. Discussion/Significance of Impact: Linking researchers with translational resources requires mutual understanding, timely communication, and coordination across teams. We systematically filled these information gaps between investigators and institutional resources. Our Virtual Concierge AI bot can help researchers navigate resources through the translational process.https://www.cambridge.org/core/product/identifier/S2059866124011336/type/journal_article |
| spellingShingle | Jonathan Gelfond Meredith Zozus Martin Goros Jennifer Potter Kimberly K. Summers Susanne Schmidt Stephanie Rowan Laura Aubree Shay 562 Mapping and navigating translational resources with generative AI Journal of Clinical and Translational Science |
| title | 562 Mapping and navigating translational resources with generative AI |
| title_full | 562 Mapping and navigating translational resources with generative AI |
| title_fullStr | 562 Mapping and navigating translational resources with generative AI |
| title_full_unstemmed | 562 Mapping and navigating translational resources with generative AI |
| title_short | 562 Mapping and navigating translational resources with generative AI |
| title_sort | 562 mapping and navigating translational resources with generative ai |
| url | https://www.cambridge.org/core/product/identifier/S2059866124011336/type/journal_article |
| work_keys_str_mv | AT jonathangelfond 562mappingandnavigatingtranslationalresourceswithgenerativeai AT meredithzozus 562mappingandnavigatingtranslationalresourceswithgenerativeai AT martingoros 562mappingandnavigatingtranslationalresourceswithgenerativeai AT jenniferpotter 562mappingandnavigatingtranslationalresourceswithgenerativeai AT kimberlyksummers 562mappingandnavigatingtranslationalresourceswithgenerativeai AT susanneschmidt 562mappingandnavigatingtranslationalresourceswithgenerativeai AT stephanierowan 562mappingandnavigatingtranslationalresourceswithgenerativeai AT lauraaubreeshay 562mappingandnavigatingtranslationalresourceswithgenerativeai |