RDAC

PROJECT SUMMARY Research and Data Analysis Core The Research and Data Analysis Core (RDAC) provides research collaboration and expertise for the UCSD LHS Center and LHS Scientists in study design, data management, data analysis, power and sample size calculations, data interpretation, and presentation of results. The resources provided by the RDAC will ensure

a centralized data management and analysis platform; a system to construct and conduct surveys, interviews, focus groups, and design workshops; and encourage, coach, and train LHS Scientists in biostatistics, clinical informatics, econometrics, and data management skills. The RDAC will be led by Dr. Michael Hogarth, a

biomedical informaticist, who will be supported by a senior biostatistician, a staff statistician, an EHR data analyst, two economists, and a qualitative researcher. A secure environment for housing EHR data will be provided by developing ephemeral enclaves that will function for the duration of the study thereby establishing

a high level of security that is HIPAA compliant. LHS Scientists will gain access through a “Virtual Research Desktops” that run inside the enclave providing tools for analysis on biomedical data, including standard statistical analysis, machine learning infrastructure, a DICOM image viewer, access to three natural language

processing platforms, and access to a Jupyter notebook styled platform (AWS SageMaker Studio) that employs intra-enclave Spark clusters for scalable computation (AWS Elastic Map Reduce). Furthermore, to address the typical latency of published research experiences and findings, we propose to quickly disseminate

key data and findings from LHS projects through a portal. This portal will be linked to LHS study metadata specified with salient terms used in the LHS literature. A Delphi method will be employed to create a standardized list for LHS projects. We plan to develop data entry forms in REDCap that adhere to our LHS

metadata model and work with LHS Scientists to enter LHS project information into this system. We will also build a LHS project portal where users can search the LHS repository for projects using the structured metadata variables, making dissemination of knowledge gleaned from LHS projects more likely to be

structured and in ‘real time.’ With this centralized hub for data management, statistical analysis, and retrieval of information on LHS projects, LHS Scientists can engage in discovery of meaningful insights, informing conclusions, and producing reports on progress toward research focused on AHRQ/PCORI priorities to support

decision-making.