- Video Repositories for Clinical Research: Overcoming Barriers and Testing Utility
- Analytics & Machine-learning for Maternal-health Interventions (AMMI): A Cross-CTSA Collaboration
- PANDA-MSD: Predictive Analytics via Networked Distributed Algorithms for Multi-System Diseases
- Perioperative Precision Medicine: Translating Science to Clinical Practice to Improve Safety and Efficacy of Opioids in Neonates, Children and Nursing Mothers
Video Repositories for Clinical Research: Overcoming Barriers and Testing Utility
Virginia Commonwealth University
Principal Investigator: Henry Rozycki, M.D.
Grant Number: 1R21TR003994-01A1
Collaborating Institution: University of Pennsylvania
Patient observation is key to provider assessment. In addition, observations are often included in clinical scoring systems, such as the Glasgow Coma Scale and the Apgar Score, that are used to classify patients and as surrogate outcomes in research — signs used in place of other measures to tell whether a treatment works. Most observation data, however, have significant problems with inter-rater reliability — the extent to which two or more people assessing information agree — which makes them less robust than objective, measurable data, such as blood test values. This unreliability is due in part to using limited, nontypical subject samples during their creation and validation. Using large data sets from a variety of sources is the recommended way to help overcome this challenge. This has led to the creation of data registries and biorepositories. To date, however, there are almost no examples of repositories that focus on observations made by video recording. Although the overall framework for the design and management of other registries and repositories is well established, concerns about privacy, security and liability risk have prevented the creation of video repositories. The primary objective of this study is to design solutions to these concerns and produce a guideline or manual for the creation of clinical video repositories.
Learn more about this project in the NIH RePORTER.
Analytics & Machine-learning for Maternal-health Interventions (AMMI): A Cross-CTSA Collaboration
The University of North Carolina at Chapel Hill
Principal Investigator: Javed Mostafa, Ph.D.
Grant Number: 1U01TR003629-01A1
Collaborating Institutions: Duke University and Wake Forest University Health Sciences
African American women across the United States experience alarmingly higher rates of maternal mortality than their white counterparts. Social determinants of health — such as education, housing, transportation and nutrition — might contribute to this disparity in maternal health outcomes, along with clinical risk factors, including high blood pressure and heart disease. However, the complex connections among these factors, along with the role they play in increasing the risk of maternal mortality, are not well understood. Further, a need exists for comprehensive health care interventions that take these combined factors into account to provide decision and communication support for patients, providers and community support workers. The Analytics and Machine-learning for Maternal-health Interventions (AMMI) initiative — a partnership among researchers at The University of North Carolina at Chapel Hill, Duke University, and Wake Forest University — aims to address these gaps by developing a machine learning–enhanced health technology framework to reduce the risk of maternal mortality in African American women.
Learn more about this project in the NIH RePORTER.
Note: This U01 was co-funded by NCATS and the NIH Office of Disease Prevention.
PANDA-MSD: Predictive Analytics via Networked Distributed Algorithms for Multi-System Diseases
University of Pennsylvania
Principal Investigator: Yong Chen, Ph.D.
Grant Number: 1U01TR003709-01A1
Collaborating Institutions: The University of Alabama at Birmingham; University of Florida; Vanderbilt University Medical Center; Global Healthy Living Foundation; Vasculitis Foundation
Researchers will develop novel ways to combine data from different sources using electronic health records from multiple CTSA hubs to create predictive models of multisystem diseases — disorders that affect multiple body systems. The project directly addresses the areas of emphasis in PAR-19-099 to “engage new collaborators in pre-existing collaborations to solve a translational science problem no one hub can solve alone.” The researchers will develop the Predictive Analytics via Networked Distributed Algorithms (PANDA) framework, which will improve risk prediction to help health care providers reach accurate diagnoses earlier. The researchers’ proposed methods directly address two major barriers: (1) lack of predictive models for multisystem conditions, and (2) lack of algorithms that effectively combine data from multiple sites in a way that preserves privacy and makes communication more efficient.
Learn more about this project in the NIH RePORTER.
Perioperative Precision Medicine: Translating Science to Clinical Practice to Improve Safety and Efficacy of Opioids in Neonates, Children and Nursing Mothers
University of Pittsburgh
Principal Investigator: Senthilkumar Sadhasivam, M.D., M.P.H.
Grant Number: 1U01TR003719-01A1
Collaborating Institutions: Johns Hopkins University; University of California, San Francisco; Indiana University; Washington University
Severe surgical pain is still poorly managed, yet clinicians must also avoid unpredictable and life-threatening opioid adverse effects, as well as long-term opioid use and misuse. Opioid adverse effects that occur around the time of surgery (perioperative) — from postsurgical nausea and vomiting to respiratory depression and death — are preventable challenges in managing surgical pain. This study aims to develop a proactive clinical practice to optimize postsurgical pain control and decrease opioid-related adverse effects. Opioid metabolism and opioids’ pain-relieving (analgesic) and adverse effects are influenced by genetic variations. Due to translational bottlenecks, lack of infrastructure and knowledge gaps in how to personalize opioid use and dose precisely for the best results, presurgical genotyping and personalized analgesia are not practiced, despite evidence, regulatory warnings, Clinical Pharmacogenetics Implementation Consortium guidelines, cost effectiveness and insurance coverage for CYP2D6 testing. Personalizing analgesia based on genetic risks will reduce opioid use and adverse effects and accelerate value-based care opportunities, particularly in children and nursing mothers. However, these opportunities are constrained by a lack of translational platforms and major gaps in our understanding of how to personalize and precisely dose opioids. This collaborative CTSA project aims to develop an innovative perioperative precision analgesia platform to reduce serious adverse outcomes of opioids and improve the safety of opioids in (1) children undergoing painful surgery and (2) nursing mothers and their infants.