Wake Forest IMPOWR Dissemination Education and Coordination Center (IDEA-CC)

点击翻译按钮获取中文摘要

PROJECT SUMMARY The HEAL Data Ecosystem is working to collect data across its projects and networks to meet FAIR (Findable, Accessible, Interoperable, Reusable) data standards. Bringing diverse data sources together will require complex data solutions to have a highly successful and accessible HEAL data commons. Meeting these data goals brings two major challenges. The first is there are existing siloed datasets that are not yet able to be combined with other data limiting the findability and accessibility. The second is collecting and organizing prospective data so that one could assure data quality and integrity that allows for interoperability and reuse of the data. To accomplish this goal, this proposal responds to the request for strategies to make data more machine learning/artificial intelligence (ML/AI ready). The focus of the parent grant is to create a research framework for the HEAL IMPOWR network and larger scientific community to harmonize combined chronic pain (CP) and opioid use disorder (OUD) data. This administrative supplement expands this mission beyond the scope of the NIH HEAL IMPOWR network to existing and future CP and OUD. The proposed work will significantly deepen and augment approaches to FAIR principles in CP and OUD data for both the HEAL network and larger NIH research community. It enhances the rigor of the parent grant by improving the larger data relevance of what we are doing beyond the NIH HEAL IMPOWR network. The long-term goal is to build a HEAL Data Ecosystem that incorporates existing data and supports the integration of prospective CP and OUD data collection. Building on our prior work, the overall objective of this project is to move CP and OUD data one step closer to FAIR by leveraging existing datasets and developing tools for new projects. The general hypothesis of the project is that leveraging existing CP & OUD data and collecting new data using ML/AI data quality standards will accelerate the impact of the HEAL Data Ecosystem. The general hypothesis will be tested by the following specific aims: (1) Transform existing dataset by mapping chronic pain/OUD CDE to demonstrate use case for making existing siloed data into a ML/AI ready format by automatically suggesting HEAL CDE annotations for already collected data based on semantic and syntactic analysis. (2) Adapt tools to support ML/AI readiness for existing and prospectively collected HEAL CDE. First, we will adapt our previously developed tools to measure and assess the semantic distance for pain/OUD CDE. This will support the development of federated transfer learning by assessing the quantitative distances using SHAP modeling of previously collected data. We hypothesize that these tools will provide infrastructure necessary to successfully develop ML/AI ready data. In aim 1, we believe that transforming existing datasets to be ML/AI ready will accelerate the harmonization of existing and prospective data for a future HEAL Data Commons. In aim 2, the development of CDE tools will support the data infrastructure quality checks to support ML/AI. The expected outcome of this project is data optimization pipelines and tools to support the goal of ML/AI ready data. The results will provide a strong basis for further development of the HEAL Data Ecosystem.