CTSA Program Webinar April 28, 2021 - clic-ctsa.org
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
CTSA Program Webinar April 28, 2021 The University of Rochester Center for Leading Innovation and Collaboration (CLIC) is the coordinating center for the Clinical and Translational Science Awards (CTSA) Program, funded by the National Center for Advancing Translational Sciences (NCATS) at the National Institutes of Health (NIH), Grant U24TR002260.
TIME TOPIC PRESENTER(S) Michael G Kurilla, MD, PhD 2:00 pm ET Welcome Director, Division of Clinical Innovation Michael G Kurilla, MD, PhD Director, Division of Clinical Innovation Joni L. Rutter, PhD 2:00 pm - 2:15 pm NCATS & CTSA Program Updates Acting Director, NCATS Clare K. Schmitt, PhD Acting Deputy Director, NCATS Boyd Knosp, MS 2:15 pm – 2:25 pm University of Iowa Enterprise Data Warehouse for Research - Phase 2 Thomas R. Campion Jr., PhD Weill Medical College of Cornell University Inclusion of Older Adults as a Model for Special Karen Bandeen Roche, PhD 2:25 pm - 2:35 pm Johns Hopkins University Populations Heidi Hanson, PhD, MS University of Utah 2:35 pm – 2:45 pm Life Course Visual Toolkit Development Shari Barkin, MD, MSHS Vanderbilt Medical Center Hongfang Liu, PhD Text Analytics Toward Semantic Interoperability and Data Mayo Clinic 2:45 pm – 2:55 pm Peter Elkin, MD, MACP, FACMI, FNYAM Sharing State University of New York at Buffalo Martin Zand, MD, PhD 2:55 pm – 3:00 pm CLIC Updates MPI, CLIC 3:00 pm ET Adjourn 2 clic-ctsa.org
NCATS/CTSA Program Updates Michael Kurilla, MD-PhD Director, Division of Clinical Innovation NCATS April 28, 2021
Budget: FY 2021 ➢ Consolidated Appropriations Act, 2021 ➢ P.L. 116-260, 12/27/2020 ➢ Contains all 12 bills funding the federal gov’t, plus coronavirus supplemental funding Dollars in millions FY 2019 FY 2020 FY 2021 Final Final Enacted NCATS Approp. 806.4 832.9 855.4 CTSA 557.8 578.1 586.8 CAN 44.5 49.1 "up to 60.0"
Budget: FY 2022 ➢ President’s FY 2022 Discretionary Request – released 4/9/21 ➢ Overview: https://www.whitehouse.gov/omb/FY-2022-Discretionary-Request/ ➢ NIH – Request for $51B ($9B increase over FY21) ➢ No details on NIH Institute/Center funding requests ➢ Detailed request coming in next few months ➢Congressional Hearings ➢ House Appropriations Subcommittee Hearing for HHS – 4/15/21 ➢ HHS Sec. Xavier Becerra was the witness ➢ No hearings for NIH scheduled yet
CTSA Program Info & Reminders NIH Policies & Updates • NOT-TR-20-036: Notice of Change to Key Dates for PAR-18-940 (CTSA U54). August 15, 2021 receipt date has been moved to July 15, 2021. • NOT-OD-21-073: Upcoming Changes to the Biographical Sketch and Other Support Format Page for Due Dates on or after May 25, 2021 • Non-Compliant Publications in RPPRs: Publications not compliant with the NIH Public Access Policy will cause a delay in review and processing of the applicable Notice of Grant Award. • NOT-TR-21-020: Key Date Change PAR-19-337 Competitive Revision Awards for the CTSA Program. Early expiration: the final September 27, 2021 due date is eliminated (there are no remaining receipt dates). POC: Pablo Cure, MD, MPH • NCATS Advisory Council (June 10-11, 2021) Several CTSA concepts will be presented: https://ncats.nih.gov/advisory/council/meetings Funding Opportunity Announcements • PAR-21-203 and NOT-TR-21-025 Clinical and Translational Science Award (CTSA) Consortium-Wide Centers: Resources for Rapid Demonstration and Dissemination (C3-R2D2) Due Date June 21, 2021 • Notice of Intent to Publish Research Opportunity Announcements for the NIH Post-Acute Sequelae of SARS-CoV-2 Infection Initiative: PASC Data Repositories and Mobile Health Platform
RADx Underserved Populations (RADx-UP) • RFA-OD-21-009 – Emergency Award: RADx-UP - Social, Ethical, and Behavioral Implications (SEBI) Research on Disparities in COVID-19 Testing among Underserved and Vulnerable Populations (U01 Clinical Trials Optional) • RFA-OD-21-008 – Emergency Awards: Community-engaged COVID-19 Testing Interventions among Underserved and Vulnerable Populations- RADx-UP Phase II (U01, Clinical Trial Optional) • NOT-OD-21-103 – Notice of Special Interest (NOSI): Emergency Competitive Revisions for NIH Grants to Add or Expand Community-engaged COVID-19 Testing Interventions among Underserved and Vulnerable Populations – RADx-UP Phase II (Emergency Supplement- Clinical Trial Optional) • NOT-OD-21-101- Notice of Special Interest (NOSI): Administrative Supplements for Rapid Acceleration of Diagnostics-Underserved Populations (RADx-UP) Phase I Projects to Address Vaccine Hesitancy and Uptake • NOT-OD-21-097 – Notice of Intent to Publish a Research Opportunity Announcement for RADx-UP Return to School Diagnostic Testing Approaches (OT2 Clinical Trial Optional) References: • https://www.nih.gov/research-training/medical-research-initiatives/radx • https://grants.nih.gov/grants/guide/COVID-Related.cfm
ACTIV-6 Re-Purposed Drugs Mild to Moderate COVID-19 X Prevent Hospitalization Clinical Coordinating Center Data Coordinating Center Site Support
Upcoming: •QA/QC Meeting: •Date: May 7, 2021 •Registration Link •CTSA Program Annual Meeting •Date: December 1-3, 2021
Working Group Reports Enterprise Data Warehouse for Research - Phase 2 Boyd Knosp, MS (University of Iowa) and Thomas R. Campion Jr., PhD (Weill Medical College of Cornell University) Inclusion of Older Adults as a Model for Special Populations Karen Bandeen Roche, PhD (Johns Hopkins University) Life Course Visual Toolkit Development Heidi Hanson, PhD, MS (University of Utah) and Shari Barkin, MD, MSHS (Vanderbilt Medical Center) Text Analytics Toward Semantic Interoperability and Data Sharing Hongfang Liu, PhD (Mayo Clinic) and Peter Elkin, MD, MACP, FACMI, FNYAM (State University of New York at Buffalo) 12 clic-ctsa.org
Enterprise Data Warehouse for Research (EDW4R) Working Group CTSA Program Webinar April 28, 2021 Boyd Knosp, MS Thomas R. Campion, Jr., Ph.D. Institute for Clinical and Translational Science
Goals and outcomes • Goals • Increase understanding of enterprise data warehouse for research (EDW4R) activities at CTSA hubs • Identify best practices to support clinical and translational researchers with electronic patient data • Outcomes • Novel studies describing EDW4R current state and next steps • References for EDW4R activities • Engagement across the CTSA community Institute for Clinical and Translational Science
Enterprise Data Warehouse for Research (EDW4R) Data Sources Transform Repositories Applications Uses Patient Data Study Self- feasibility EHR service Databases Normalize Pop. health Researchers External Data marts data Basic informatics Filter Registries Query Quality improvement tools Internal data Research networks
EDW4R Working Group Timeline • Spring 2018: initial discussion at iEC (iDTF) meeting • CY 2019: Phase 1 • CY 2020: Phase 2 • Spring 2021: planning for Phase 3 Institute for Clinical and Translational Science
EDW4R Phase 1: Study of CTSA Hubs • Focus areas • Architecture • Access and engagement • Service management • Maturity • Qualitative methods • 20 semi-structured interviews with CTSA informatics leaders • Directed content analysis Institute for Clinical and Translational Science
EDW4R Phase 1: Academic Output Understanding enterprise data warehouses to support clinical and translational research Thomas R Campion, Jr, Catherine K Craven, David A Dorr, Boyd M Knosp Journal of the American Medical Informatics Association, ocaa089, https://doi.org/10.1093/jamia/ocaa089 Published: 17 July 2020 Institute for Clinical and Translational Science
EDW4R Phase 2: Follow-up Study • Focus areas (identified in Phase 1 for future research) • Enterprise IT relationship • Data governance • Workforce • Cloud computing • Qualitative methods • Maturity index pilot: https://redcap.link/nfu8wswl Institute for Clinical and Translational Science
EDW4R Phase 2: Community Discussions • April: COVID-19 support from EDW4R • May: Cloud with NIH STRIDES project leaders • June: Data governance part 1 • July: Data governance part 2 (with Adam Wilcox of CD2H) • August: Work force & JAMIA publication • September: Relationship with enterprise IT • October: Maturity model • November: UT Health honest broker approach with Elmer Bernstam • December: Stanford cloud computing approach with Michael Hallas • February 2021: Phase 2 analysis discussion & maturity model pilot kick off Institute for Clinical and Translational Science
EDW4R Phase 2: Deliverables • Conference abstracts in 2020 (cancelled) • AAMC GIR • AMIA Summits • Participation in CD2H maturity model project • AMIA Informatics Summit 2021 • Panel on EDW4R operations and trends • Poster on EDW4R maturity • Translational Science 2021 • Panel on maturity models • AMIA Annual Symposium 2021 • Podium abstract on EDW4R phase 2 (under review) Institute for Clinical and Translational Science
EDW4R: JAMIA Special Focus Issue Call for Papers for Special Focus Issue Best Practices in Research Patient Data Repositories Guest Editor-in-Chief: Leslie A. Lenert, Medical University of South Carolina Submissions due Focus Issue Guest Editors: Shawn N. Murphy, Mass General Brigham June 15, 2021! Michael J. Becich, University of Pittsburgh School of Medicine Thomas R Campion, Weill Cornell Medical College Boyd M. Knosp, University of Iowa Genevieve Melton-Meaux, University of Minnesota Shyam Visweswaran, University of Pittsburgh School of Medicine
Best Practices in Research Patient Data Repositories (RPDRs) –Topics of Interest • Gathering leadership support and funding for • Novel Academic Industry Partnerships for the RPDRs RPDRs • Showing a return on investment for the RPDRs • Data governance (all flavors) • Novel architectures (including cloud • Workforce for the RPDRs environments) for RPDRs • The role of enterprise IT and/or research • Design of “ideal” next-generation RPDRs faculty in RPDR operations; intersection of RPDR with academic informatics; intersection • RDW assessment tools such as maturity of RPDR with quality of care models and data quality processes • Organization, data structures and ontologies, • Lessons learned from implementation of tools to access, for RPDRs RPDRs • The role of RPDRs in learning health systems • Relationship of RPDRs to Enterprise Data Warehouses and Clinical Data Warehouses • Security and compliance practices for RPDRs • Innovative methods for supporting RPDRs • RPDR literacy for researchers • Sustainability models for supporting RPDRs
EDW4R Phase 3: Planning • Step back consulting • Best practice products • Questions PIs should ask about enterprise IT • Seven things to know about EDW4R and cloud computing • Maturity index • Case studies • Natural language processing and EDW4R • Sustainability approaches Institute for Clinical and Translational Science
Questions and comments • EDW4R Phase 3 interest: CLIC coordinators/website • Boyd Knosp: boyd-knosp@uiowa.edu • Tom Campion: thc2015@med.cornell.edu Institute for Clinical and Translational Science
Inclusion of Older Adults as a Model for Special Populations Karen Bandeen-Roche Frank Hurley and Catharine Dorrier Professor and Chair Department of Biostatistics On Behalf of the Workgroup CTSA Program Meeting April 28, 2021
Goals and Outcomes • Promote the integration of special populations in across the human lifespan • Train and cultivate the translational science workforce • Innovate processes to increase the quality and efficiency of translational research • Implementation of a process or best practice • Knowledge dissemination relevant to CTSA goals • Increased collaboration © 2014, Johns Hopkins University. All rights reserved.
Workgroup Members • Karen B-R • Marco Pahor, UFL • Cynthia Boyd, JHU • Todd Manini, UFL • Elizabeth Eckstrom, OHSU • Steve Anton, UFL • Steve Kritchevsky, WFU • Daniel Mullins, UMD • Susan Stark, WUSL • Wendy Kohrt, Colorado • Jay Magaziner, UMD • Kady Nearing, Colorado • Elena Volpi, UTMB • Jerry Gurwitz, Meyers Inst. • Mark Supiano, Utah Enterprise Committee Affiliation: LIFESPAN © 2014, Johns Hopkins University. All rights reserved.
Why is the work needed? Policy demands it “Applications and proposals involving human subjects research must address plans for including individuals across the lifespan in the PHS Human Subjects and Clinical Trial Information Form. Any age- related exclusions must include a rationale and justification based on a scientific or ethical basis.” (NIH Policy on Inclusion across the lifespan effective January 25, 2019) “For most cancers, clinical trials should include a representative population of older adults. Older adults, including those with frailty, should be enrolled in all phases of clinical trials, when they can be safely and ethically enrolled.” (FDA Draft Guidance or Industry Sponsored Trials in Oncology, March 2020). © 2014, Johns Hopkins University. All rights reserved.
Barriers to Representation of Older Adults in Cardiovascular Disease Trials Before and After the Inclusion Across the Lifespan Policy Nanna MG. JAMA Intern Med. 2020 3 © 2014, Johns Hopkins University. All rights reserved. 0
Inclusion of Older Adults . . . • Meets the Needs of a Changing Nation • Satisfies Our Desire for Justice • Is Consistent with Our Scientific Values • Has Practical Benefits • Is Consistent with the Flow of Policy clic-ctsa.org
Workgroup Goals and Deliverables 1. Conference on Older Adult Inclusion > Identify knowledge / practice gaps © 2014, Johns Hopkins University. All rights reserved.
Research Centers Coordinating Network / Workgroup Workshop: February 22-23, 2021 © 2014, Johns Hopkins University. All rights reserved.
Research Centers Coordinating Network / Workgroup Workshop: February 22-23, 2021 • 15 speakers spanning 14 institutions • Sessions • Current Situation: In subspecialties • Barriers • Making the case • Next steps: Implementation! • Follow ups: Implementation roundtable, grants Website: https://www.afar.org/events/rccn-event- inclusion-of-older-adults-in-clinical-research © 2014, Johns Hopkins University. All rights reserved.
Workgroup Goals and Deliverables 1. Conference on Older Adult Inclusion ➢ Identify knowledge / practice gaps 2. Toolkit: Modular Presentation Materials ➢ Workshops; Advocacy; Knowledge ➢ Reference repository © 2014, Johns Hopkins University. All rights reserved.
Modules Why include older adults (You should!) Practical strategies Consenting/Inclusion considering cognitive impairment Multimorbidity Polypharmacy Design and data analysis strategies Focusing protocols on geriatric outcomes Community engagement Cultural considerations / diverse inclusion Institutional socialization © 2014, Johns Hopkins University. All rights reserved.
Workgroup Goals and Deliverables 1. Conference on Older Adult Inclusion ➢ Identify knowledge / practice gaps 2. Toolkit: Modular Presentation Materials ➢ Workshops; Advocacy; Knowledge ➢ Reference repository 3. Pilot Workshops ➢ ACTS; SGIM; Subspecialty; AGS © 2014, Johns Hopkins University. All rights reserved.
You Should be Recruiting Older Adults Stephen B. Kritchevsky, PhD Wake Forest School of Medicine & Sticht Center for Healthy Aging and Alzheimer’s Prevention clic-ctsa.org
Workgroup Goals and Deliverables 1. Conference on Older Adult Inclusion ➢ Identify knowledge / practice gaps 2. Toolkit: Modular Presentation Materials ➢ Workshops; Advocacy; Knowledge ➢ Reference repository 3. Pilot Workshops ➢ ACTS; SGIM; Subspecialty; AGS 4. Inclusion Tracking Data Template 5. Lessons for other special populations © 2014, Johns Hopkins University. All rights reserved.
How you can get involved • Join us! • Karen Bandeen-Roche - kbandee1@jhu.edu • Advocate for older adult inclusion • Modular materials to be available soon • CTSA appeal: RFAs, further institutional and implementation resources • Next steps • Finalize, polish slide set • Adjunct materials • 1-2 more workshops • Tracking data template, lessons beyond older age © 2014, Johns Hopkins University. All rights reserved.
Life Course Research Visual Toolkit: Why and What Heidi Hanson, PhD, MS Assistant Professor of Surgery University of Utah Shari Barkin, MD, MSHS Professor of Pediatrics Vanderbilt University Medical Center
Developmental Origins of Health and Disease: A Lifecourse approach to the prevention of non- communicable disease. Baird J et al. Healthcare, 2017.
Traditional approaches to science • Based on a Newtonian paradigm • Linear cause and effect
Past Approaches: Simplify Problem: It ignores the interdependence of the factors. We are understanding health and disease in a vacuum.
COMPLEXITY SCIENCE 4/28/2021 45
46
Hanson, HA., Leiser, C., Bandoli, G., Pollock, B., Karagas, M., Armstrong, D., . . . Barkin, S. (2021). Charting the life course: Emerging opportunities to advance scientific approaches using life course research. Journal of Clinical and Translational Science, 5(1), E9. doi:10.1017/cts.2020.492
Introducing…… • 6 full webinars with more than 15 national and international experts • 32 number of shorter educational skills building videos that include content and, importantly, methods for how to implement this research 48
•An Introduction to Life Course Research and Complexity Science •The Promise of the Exposome: Scaling it Up •All of Us Research Program •Understanding Health Across the Lifecourse •Journey into Complexity Science: The Promise of the Exposome •An Overview of Longitudinal Trajectory Methods •The Importance of Measuring Social and Chemical Stressors •Using Electronic health Records for Medical Research • New Methods for Identifying Complex Patterns of Disease in •Discovery of a Drug-Drug Interaction Families and Linking them to their Etiological Roots •A How-to Guide for Studying the Elusive Exposome in Complex • Assessing Multiple Exposures Across Time Disease with Large Data • Machine Learning to Classify Subgroups of Disease and •Dissecting P=G+E with Big Data Predict Outcomes •Applying Data Science Techniques to Identify Phenotypic • Developing and Validating Ways to Model High-Dimensional Variation Data •Understanding Potential Biases Inherent in Data •Linkage Across Diverse Data Resources •Project Viva: Challenges and Lessons Learned •Utilizing Data Science Resources to Prepare and Package •Research Opportunities with a Single Cohort and Beyond Integrated Datasets •How to Build a Longitudinal Cohort •Using Trajectory Methods to Identify Sensitive Periods During •Using Microsimulations to conduct Virtual Experiments Pregnancy •How to Implement a Microsimulation Model: Application • Using Trajectory Methods to Capture Timing, Dose, and Duration of Exposure •Evaluating Interventions with a Complex Systems Approach •Using Trajectory Methods to Identify Sensitive Periods During •Moving Towards a Complex Systems Approach to Health Pregnancy Intervention Research • Using Trajectory Methods to Capture Timing, Dose, and Duration of Exposure
4/28/2021 50
Dissemination Strategy Target audience: Learners at all stages interested in understanding the importance of lifecourse research, the potential applications, and emerging methods to conduct this type of research Current dissemination strategy: • CLIC website and blog • Highlight on CTSI’s webpages • Distribute to academic societies reaching across the lifespan • Send to the leaders of medical schools and schools of public health 51
Working group members Heidi A. Hanson,Shari Barkin, MD, Frederick Kaskel, Joemy Ramsay, Nate O’Neil, MC PhD, MS MSHS MD, PhD PhD, MS University of Utah University of UtahVanderbilt Medical Albert Einstein University of Utah Center School of Medicine Margaret R. Karagas Dartmouth College Jonathan N. Tobin Albert Einstein College of Medicine Carrie Dykes University of Rochester Maureen Monaghan Children's National Peter Szilagyi University of California, Los Angeles Christopher Seplaki University of Rochester Elizabeth Eckstrom Oregon Health and Science University Mark Schleiss University of Minnesota Twin Cities Gretchen Bandoli University of California San Diego Bradley Pollock University Of California Davis Anne Hoen Dartmouth College
4/28/2021 53
WG: Text analytics towards semantic interoperability and data sharing for the research use of EHR Progress Report to iEC Lead Team Peter Elkin, University at Buffalo Hongfang Liu, Mayo Clinic Justin Guinney, Sage BioNetworks clic-ctsa.org
Text Analytics WG Goals • Goal 1: Containerize open source text analytics with instructions for their integration into local applications • Goal 2: Contribute to the community-wide effort on data sharing through collaboration with CD2H and TIN • Goal 3: Create a community share task to advance text analytics for concept encoding and de-identification • Goal 4: Load de-identification and text analytics applications to the NCATS cloud clic-ctsa.org
Text Analytics WG Deliverables Year 1 • NLP Cloud Sandbox Pilot de-identification of narratives (led by Bradley W. Taylor & Russ Waitman) • OHNLP Artifact Discovery and Preparation Toolkit, NLP-ADAPT (led by Serguei Pakhomov) (github.com/nlpie/nlp-adapt-kube) – see Best Practice Presentation at Monday, September 16, 2019. • Ongoing Text Analytics WG and CD2H collaboration for N3C NLP Effort • An End-to-End Sample Implementation of NLP for Multi-site Cohort Studies • Catalog best practice and review articles clic-ctsa.org
Text Analytics WG Deliverables Year 2 • Text Analytics WG and CD2H collaboration for N3C NLP Effort and NLP Sandbox – in production • A text analytics infrastructure and • A community-wide federated NLP evaluation solution • Codification of Clinical Problems in SNOMED CT, Lab test result names in LOINC and Medications in RxNorm • Catalog best practice and review articles – in progress clic-ctsa.org
U01TR002062 Open Health Natural Language Processing Clinical and Translational Science Collaboratory Institute (University of Minnesota) Center for Clinical and Translational Science (Mayo Clinic) Irving Institute for Clinical and Translational Research (Columbia University) Partnership Center for Clinical and Translational Team Sciences (UT Health) University at Buffalo Aim 1: Obtain NLP NLP Aim 3: Develop privacy- artifacts across sites Innovations preserving computational for improving NLP phenotyping enhanced algorithm development with NLP Aim 2: Generate a Aim 4: Partner with synthetic text corpus diverse communities for for exploratory analysis Privacy- translation science Computational of clinical narratives Preserving Phenotyping excellence leveraging Computing EHR clic-ctsa.org
National COVID Cohort Collaborative https://github.com/OHNLP/N3C-NLP-Documentation/wiki clic-ctsa.org
View N3C NLP as a Real-world AI Project • Behave as it • Can provide should be evidences of the decision Health Explainability Security Reproducibility • Can tolerate attacks, i.e., • All predictions efforts to must be change or reproducible manipulate its behavior Reproducibility Explainability Implementability Interoperability Transparency clic-ctsa.org
History of Technology Stack of MVP N3C NLP ● History ○ 2009 - BioTagger GM [1] and 2011 - Concept extraction [2] ■ Dictionary Lookup and Machine Learning (Weak Supervision) ○ 2013 - Refactor to UIMA and develop OHNLP common type systems for facilitating interoperability ○ 2013 - MedTaggerIE [3], clinical researchers want explainable NLP (Human-AI Trust) ○ 2016 - Big Data NLP Infrastructure [5] ○ 2019 - MedTagger Suite to Enterprise Implementation - NLPaaS [6] 1. Torii M, Hu Z, Wu CH, Liu H. BioTagger-GM: a gene/protein name recognition system. J Am Med Inform Assoc. 2009;16(2):247-255. doi:10.1197/jamia.M2844 2. Torii M, Wagholikar K, Liu H. Using machine learning for concept extraction on clinical documents from multiple data sources. J Am Med Inform Assoc . Sep-Oct 2011;18(5):580-7. doi: 10.1136/amiajnl-2011-000155. Epub 2011 Jun 27. 1. Wu S, Kaggal C, Dligach D, et al A common type system for clinical natural language processing. J Biomed Semantics . 2013 Jan 3;4(1):1. doi: 10.1186/2041-1480- 4-1. 2. Liu H, Bielinski SJ, Sohn S, et al. An information extraction framework for cohort identification using electronic health records. AMIA Jt Summits Transl Sci Proc. 2013;2013:149-153. Published 2013 Mar 18. 3. Kaggal V, Komandur R, Mehrabi S, et al.Toward a Learning Health-care System - Knowledge Delivery at the Point of Care Empowered by Big Data and NLP Biomed Inform Insights . 2016 Jun 23;8(Suppl 1):13-22. doi: 10.4137/BII.S37977. eCollection 2016. 4. Wen A, Fu S, Moon S, et al. Desiderata for delivering NLP to accelerate healthcare AI advancement and a Mayo Clinic NLP-as-a-service implementation. NPJ Digit Med. 2019;2:130. Published 2019 Dec 17. doi:10.1038/s41746-019-0208-8 clic-ctsa.org
NLP Algorithm Development Process Not surprising to everyone, majority of the effort in clinical concept extraction application is “Task formulation” and “Annotation guideline development” Sunyang Fu et al. Development of Clinical Concept Extraction Applications: A Methodology Review, Journal of Biomedical Informatics (2020): 103526. clic-ctsa.org 62
The portability of the statistical NLP system (i.e., Generalizability Challenges machine learning-based system) highly depends on the distributional statistics. The institutional, EHR, and documentation variations across different institutions make the underlying distribution dramatically varying. Training a statistical NLP system in a new environment is very expensive. clic-ctsa.org
NLP Algorithm Implementation Environment To innovate, domain expertise must be collected and preserved To facilitate, toolsets should engage and empower domain experts To accelerate, NLP platforms should be responsive and scalable clic-ctsa.org
N3C Text Analytics Backbone Infrastructure clic-ctsa.org
N3C NLP Open Development Framework Due to PHI concerns, NO large clinical document collections. If we cannot share the text data with the public, can we share NLP knowledge (lexicon, pattern)? clic-ctsa.org
Ongoing: Incorporate Information Retrieval into the infrastructure https://medinform.jmir.org/2020/10/e17376 clic-ctsa.org
N3C NLP Community Participatory Progress Updates ○ Mayo Clinic (Hongfang Liu) ○ Tufts Medical Center (Andrew Williams, Robert Miller) ○ University of Kentucky (Ramakanth Kavuluru, Daniel Harris) ○ University of Kansas Medical Center (Mei Liu) ○ University of Minnesota (Rui Zhang) ○ Wake Forest University (Umit Topaloglu) ○ University of Washington (Meliha Yetisgen) ○ University of Alabama at Birmingham (James Cimino, John Osborne) ○ John Hopkins University (Chris Chute, Masoud Rouhizadeh) ○ Columbia University (Karthik Natarajan) ○ University of Buffalo (Peter Elkin) ○ Stony Brook (Joel Saltz, Janos Hajagos) ○ UT Health (Hua Xu, Kirk Roberts) ○ University of North Carolina (Emily Pfaff) ○ University of New Mexico (Larissa Myaskovsky) ○ …. clic-ctsa.org
CD2H NLPSandbox.io: Overcome data access barriers in biomedical tool benchmarking ● Critical patient information derived from academic research, health care and clinical trials are off limit for traditional data-to-model challenges. Existing barriers include: ○ Access to big and sensitive data ○ Lack of effective frameworks for assessing performance & generalizability clic-ctsa.org
NLPSandbox.io: Benchmarking NLP tools on private data clic-ctsa.org
PHI Annotation & De-identification first series of NLP Sandbox tasks ● The first series of tools that can be benchmarked on NLPSandbox.io target PHI Annotation & De-identification in clinical notes. ○ Date annotation ○ Person name annotation ○ Physical address annotation ● The community can contributes by submitting new NLP Sandbox tasks NLP Sandbox tool specifications Annotation Knowledge graph generation nlpsandbox/nlpsandbox-schemas (GitHub) clic-ctsa.org
Development of NLP Sandbox tools ● Robust and reliable ● Reproducible ● Re-usable ● Portable ● Cloud-friendly nlpsandbox/date-annotator- example nlpsandbox/nlpsandbox-schemas nlpsandbox/date-annotator-example- java clic-ctsa.org
“The tool deployed in production is a tool benchmarked” ● A small change made to a tool when adapting it to a production environment can dramatically reduce its performance. The NLP Sandbox promotes the development of tools that are production-ready, interchangeable, and re-usable. The performance of the NLP Sandbox PHI Deidentifier improves as more performant building blocks are submitted to NLPSandbox.io. phi-deidentifier.nlpsandbox.io clic-ctsa.org
NLPSandbox.io Join us on the NLP Sandbox Discord Server nlpsandbox.io/discord nlpsandbox.io GitHub: https://github.com/nlpsandbox Data hosting sites: ● Sage Bionetworks (2014 i2b2 de-identification challenge data) ● Medical College of Wisconsin (private data - onboarding) ● Mayo Clinic (private data - onboarding) clic-ctsa.org
Record Process Collaborative Pipeline Web Server Enterprise Enterprise Transcribed Record Java Bean Container Processed Record Record to Processor to Storage Map Text To and Return Map Text To Terminology and Terminology Store EMR / OMOP EMR Terminology Terminology Repository Server Server Intelligent Query Intelligent Query to Database Database Handle Query Handle Queryand and Explode Matches Explode Matches Query to Processor and Return Query Record Retrieval Process clic-ctsa.org
clic-ctsa.org
NLP Component:: Compositional Expression Generator clic-ctsa.org
UIMA NLP & CE Example clic-ctsa.org
clic-ctsa.org
"DOC_ID SEC_ID PROP_ID CONCEPT_CODE INEX_TYPE PTS" "2 52 1 23685000 7 1" "2 52 1 255302009 7 1" "2 52 1 79619009 7 1" "2 52 2 53059001 7 1" "2 52 2 237679004 7 1" "2 52 2 11092001 7 1" "2 52 3 49436004 7 1" "2 51 13 43364001 7 1" "2 51 14 258707000 7 1" "2 51 14 255260001 7 1" "2 51 14 102522009 7 1" "2 51 15 226630009 7 1" "2 51 15 51440002 7 1" "2 51 15 182334003 7 1" "2 51 15 43364001 7 1" "2 56 43 229799001 7 1" "2 56 43 clic-ctsa.org 258684004 7 1"
Observational Data are formatted for OMOP (OHDSI) and i2b2,and PCORNet clic-ctsa.org
clic-ctsa.org
Ensemble Machine Learning Method clic-ctsa.org
Ongoing Activities Towards network-wide text analytics infrastructure deployment ➢Pilot testing with four N3C sites, each adopts a different common data model ➢For N3C data ingestion and harmonization, collaboration with the Standard Operation Procedure(SOP) will be posted on the N3C Phenotype Data Acquisition wiki: https://github.com/National-COVID-Cohort- Collaborative/Phenotype_Data_Acquisition/wiki/NLP-Submission-Process Bringing model to data for federated benchmarking and seamless NLP deployment ➢Example COVID 19 rulesets will be posted on NLPSandbox • NLPSandbox is currently running a deid sharetask • https://nlpsandbox.io clic-ctsa.org
Ongoing: Catalog review and best practice articles • iEC Text Analytics WG monthly meeting: the best practice session • Kirk Roberts from UT Health – Dec 21, 2020 • Meliha Yetisgen from University of Washington – Jan 18, 2021 • Stephane Meystre from MUSC – March 15, 2021 • Working with CD2H on a scope review related to the use of NLP for clinical and translational research with standard operation procedures (SOPs) clic-ctsa.org
Thank you clic-ctsa.org
CLIC Updates 87 clic-ctsa.org
CLIC Career Development Community • An interactive space available to the CTSA Program and greater translational science community where individuals can collaborate, network, and share ideas with peers from across the consortium. • Who Can Join? Anyone within the translational science community interested in the career development of trainees, scholars, and CRPs. • Questions? education@clic-ctsa.org Join Now! https://clic-ctsa.org/career-development-community 88 clic-ctsa.org
Exploring the Inclusion of Community Hospitals in Clinical Research Thursday, May 27, 2021 11:00 AM – 3:00 PM ET Registration Now Open! 89 clic-ctsa.org
CTSA Program Working Group Application Submission - Cycle 6 • What is required in the application? ‒ Individuals/Groups need to propose and deliver well-defined projects or deliverables that fill identified translational gaps and/or further the CTSA Program objectives in high priority areas within clinical and translational science. • Who can submit an application? Application Deadline: ‒ Any group or individual within the CTSA Program Tuesday, June 1, 2021 (no later than 11:59 pm EST) • Who will receive CLIC support? ‒ Any Working Group approved by the CTSA Program Steering Committee • How does someone apply? ‒ Log onto the CLIC website under the Groups; Working Group and then Working Group Applications. clic-ctsa.org
Transition to CM-PRISM Software Platform • Scorecard licenses expire April 30, 2021 • Hubs wishing to continue using Scorecard for non-Common Metrics data – contact Clear Impact • Migration of Careers and Informatics data • Pilot funding and IRB not supported on new platform • “How to Use CM-PRISM” webinars • Thursday, May 6th at 2:00 ET • Monday, May 17th at 3:00 ET • Visit the CMI webpage for details • 2020 CMI data due August 31, 2021 – submitted in CM-PRISM 91 clic-ctsa.org
Insights to Inspire 2021 – The Basics Informatics: The Journey to Interoperability • Series of webcasts • 12 webcasts • 10-20 minutes in length • Based on themes identified in 2019 TTC Plans • Subject Matter Experts present on topics • Dissemination • Posted on CLIC website • Posted on YouTube and Vimeo • Launch Date – June 2021 • Stay tune for additional details 92 clic-ctsa.org
Reminders • The next scheduled CTSA Program Webinar: ▪ Date: May 26, 2021 ▪ Time: 2:00-3:00pm Eastern Time. • Group coordinator: ▪ ctsa_program_webinar@clic-ctsa.org • CTSA Program Webinar meeting archives/information: ▪ https://clic-ctsa.org/groups/ctsa-program-group 93 clic-ctsa.org
CLIC/NCATS Communication Channels Sharing Content: CLIC Website Twitter ❖ News (Consortium News & Mike’s Blog): ❖ NCATS: twitter.com/ncats_nih_gov clic-ctsa.org/news ❖ CLIC: twitter.com/CLIC_CTSA ❖ Events: clic-ctsa.org/events ❖ Hashtag: #CTSAProgram ❖ Education & Career Development Gateway: https://clic-ctsa.org/education-careers ▪ Education Clearinghouse: clic-ctsa.org/education CLIC Contact Us ▪ Opportunities Board: https://clic- ❖ Have a question and not sure where to direct it? ctsa.org/opportunities-board clic-ctsa.org/contact ▪ Diamond: https://clic-ctsa.org/diamond ❖ Funding opportunities (RFAs): Newsletters ▪ Synergy Papers: ❖ CTSA Ansible: Subscribe Here https://clic-ctsa.org/collaboration/clic-synergy-papers ❖ CLIC News Roundup: Subscribe Here ▪ Un-Meeting https://clic-ctsa.org/collaboration/clic-un-meetings ❖ NCATS e-Newsletter: ncats.nih.gov/enews 94 clic-ctsa.org
CTSA Program Initiative Channels ACT Network Recruitment Innovation Center (RIC) ❖ Website: http://www.actnetwork.us/National ❖ Website: ❖ Subscribe to newsletter: https://trialinnovationnetwork.org/recruitment- https://bit.ly/2HQGsM5 innovation-center ❖ Subscribe to newsletter: https://bit.ly/2OpEDHc IREx SMART IRB ❖ Website: https://www.irbexchange.org ❖ Website: https://smartirb.org ❖ Subscribe to newsletter: https://bit.ly/2TtQG7b ❖ Subscribe to newsletter: https://bit.ly/2JFbiK3 National Center for Data to Health Trial Innovation Network (TIN) (CD2H) ❖ Website: https://trialinnovationnetwork.org ❖ Website: https://ctsa.ncats.nih.gov/cd2h/ ❖ Subscribe to newsletter: https://bit.ly/2TXHQDZ ❖ To Join: N3C & CD2H Login/Registration 95 clic-ctsa.org
You can also read