Having successfully established the infrastructure and governance for cohort discovery spanning multiple regional partners, and ways to provision research data, the Biomedical Informatics Program (BIP) looks to the next logical step: helping people use contemporary computational methods on data to improve healthcare. Tackling this challenge requires us to engage the continuum of stakeholders (researchers, clinicians, policy makers), educating as well as offering new services enabling agile, data-driven approaches to be tested and rolled out.
To achieve its objective, BIP has three specific aims.
Aim 1. Enhance researchers’ access to high-quality patient data. BIP will further investigators’ ability to identify research cohorts, ensuring querying/sharing for NCATS programs, including Accrual to Clinical Trials (ACT) and the Trial Innovation Network (TIN); and make available new data types. Across our Hub we will evolve secure environments and tools for linkage, discovery and navigation of data alongside organized CTSI-wide expertise supporting its use.
Aim 2. Facilitate reproducible data-driven science. The National Academy of Sciences, Engineering, and Medicine recently discussed the problem of biomedical scientific reproducibility and replicability, with implications for DS/AI computational approaches. We will develop an informatics platform and governance policies to foster sharing and reproducibility of data-driven methods and analyses.
Aim 3. Translate data science (DS) and artificial intelligence (AI) into practice. We will advance the use of these computational methods by coalescing interdisciplinary expertise and by establishing active testbeds for projects, addressing the “last mile” of translation and socio-technological issues through implementation science.
Developments and experiences from Aims 1-3 will be shared with CD2H and aligned with NIH initiatives, including uptake of FAIR and HL7 FHIR standards. Importantly, BIP’s efforts integrate across all CTSI Programs.