Leveraging the All of Us Database for Primary Care Research with Large Datasets

The National Institutes of Health (NIH) are supporting the All of Us research program, a large multicenter initiative to accelerate precision medicine. The All of Us database contains information on greater than 400,000 individuals spanning thousands of medical conditions, drug exposure types, and laboratory test types. These data can be correlated with genomic information and with survey data on social and environmental factors which influence health. A core principle of the All of Us program is that participants should reflect the diversity present in the United States population.The All of Us database has advanced many areas of medicine but is currently underutilized by primary care and public health researchers. In this Special Communication article, I seek to reduce the "barrier to entry" for primary care researchers to develop new projects within the All of Us Researcher Workbench. This Special Communication discusses (1) obtaining access to the database, (2) using the database securely and responsibly, (3) the key design concepts of the Researcher Workbench, and (4) details of data set extraction and analysis in the cloud computing environment. Fully documented, tutorial R statistical programming language and Python programs are provided alongside this article, which researchers may freely adapt under the open-source MIT license. The primary care research community should use the All of Us database to accelerate innovation in primary care research, make epidemiologic discoveries, promote community health, and further the infrastructure-building strategic priority of the family medicine 2024 to 2030 National Research Strategy.