The Kaiser Permanente Research Bank as a Genomic Data Resource

Importance: Large biobanks have become significant engines of discovery-based health research; the collection of large-scale genetic and other molecular data from a deeply phenotyped cohort enables epidemiological studies, biomarker discovery, predictive modeling, and health economic research. In many biobanks, participant data collection is cross-sectional and relies on self-reported medical histories. Furthermore, biobanks may lack ancestral diversity and may need to be enriched for healthy individuals with lower rates of chronic disease. Objective: To describe the scope and purpose of the Kaiser Permanente Research Bank (KPRB), a large biobank created from medical records, health and socioeconomic surveys, genomic data, and biospecimens of the Kaiser Permanente patient population. Observations: The KPRB was intended to compile longitudinal medical records data and build a biospecimen collection that included a patient population that was wide and diverse, enabling a wide variety of research pursuits. KPRB participant medical records, survey data, and biospecimens were compiled to enable researcher use. Genomic DNA from participant samples were genotyped and these data were used in genome-wide association studies of blood pressure phenotypes. The KPRB achieved its goal of recruiting and consenting 400 000 participants (449 122 in total) and making available to the broader research community the participant data collected. The KPRB's comprehensive medical records extend retrospectively up to 28 years (median [IQR], 20.9 [14.0-28.7] years) and continue to accumulate patient data for as long as a participant is a KP member. KPRB participants represent a diversity of race and ethnicity groups (38 209 Asian [8.5%], 26 106 Black [5.8%], 50 661 Hispanic [11.3%], 317 630 White [70.7%]) and health outcomes, including both prevalent disease at the time of recruitment and incident disease diagnosed postrecruitment, that are typical of US-based populations. Genomic data collected on nearly all participants has been quality controlled and subjected to imputation. A new genome-wide association study (GWAS) analysis of blood pressure phenotypes in all genotyped KPRB participants has been conducted to validate previously published associations and confirm the quality of both the genomic and medical records data in the KPRB. Conclusions and Relevance: The KPRB, with genetic data generated on a large and diverse cohort with comprehensive medical records, survey data, and biospecimens, will support the growing interest in large biobank-based research and enable new avenues of research. GWAS analyses of common phenotypes measured as part of routine medical care indicated that KPRB is a robust resource linking health outcomes with genetic data that will yield valuable research findings. Access to the KPRB resource is available to external researchers with compelling research proposals through KPRB's Access Review Committee.