A current participant in the eMERGE-II consortium, Columbia serves a racially and ethnically diverse patient population in New York City, and has a strong tradition of community engagement. We have made significant contributions to the goals of eMERGE-II, including developing and evaluating electronic health records-based phenotyping algorithms; understanding data biases, data missingness, and other data quality issues in EHR data and their impact on phenotyping; defining a research agenda for next-generation EHR phenotyping; exploring the use of patient self-reported health status data to complement EHR data for phenotyping; developing novel methods for hereditability estimation; designing informatics interventions to integrate patient care and clinical research workflows and to link EHR and sequence data with genomic knowledge for decision support; communicating genetic risk to patients; addressing patients' preferences for returning incidental findings; and investigating the impact of returning results on patients and clinicians. Columbia has also established Precision Medicine as a major university-wide initiative. To date, our biobank has accumulated a multiethnic cohort of 26,310 individuals with their samples linked to our EHR data, among which we currently have exome sequence data on 3,059 patients and consent for broad genetic discoveries and wide data sharing without re-consent from 7,648 patients. This includes nearly 4,000 patients with rich self- reported health status information, who are representative of the Northern Manhattan community, and were not pre-selected based on any specific disease or diagnosis. Our proposal for eMERGE-III builds on our prior work and expertise in genomic medicine. Our four specific aims will be accomplished by wide dissemination of data and phenotyping algorithms, close collaboration with eMERGE and other research consortia (e.g., CSER, LEGACY, DHEAMS, OHDSI, CTSA, PCORI, and so on), and by using standards-based formal methods.
Aim 1 : Advance next-generation phenotyping by designing, validating, and sharing high-throughput, data quality-aware, standards-based phenotyping methods.
Aim 2 : Perform genetic association studies of rare variants with diverse clinical phenotypes through broad collaboration with the eMERGE network and other phenotyping research communities.
Aim 3 : Develop practical, scalable learning mechanisms for returning results by leveraging a genomic patient portal and genetic providers to dynamically elicit and incorporate patient preferences for return of genomic results, returning results, and studying patient understanding of returned results.
Aim 4 : Provide genomic decision support by enhancing and validating our clinical and informatics infrastructure for genomic decision support with learning mechanisms for tailored shared decision-making.
This project addresses the imperative needs of patients and clinicians for using genomic knowledge for disease prevention and health improvement. We will develop scalable methods for next generation phenotyping using EHR data, standards-based computable and portable phenotype knowledge, technology for embedding current evidence from the literature to provide genomic medicine decision support to clinicians at the point of care, methods for providing alerts to patients with genetic disease risks through genomic patient portals, and novel models for obtaining patient preferences and for returning genomic results to patients.
Showing the most recent 10 out of 32 publications
