11 a.m. - noon Location: FN 2.102
******NOTE THE CHANGE IN DATE******
We predict that the novel assembly of data structures that make use of data that exists in electronic health records and clinical laboratory information systems can significantly outperform prevailing statistics in the modalities of clinical trial administration and patient recruitment. What’s more, given the ever increasing trend of individualized medicine, we believe that augmenting these datasets with deep exon sequencing will provide the granularity of patient attributes that can more extensively inform patient inclusion and exclusion criteria in clinical trials of the future. Seventy percent of the data that exists in electronic health records is comprised of clinical laboratory generated data, depending on the specialty/sub specialty of the provider. Very little of this data exists in any kind of structure because data transmission from laboratory information systems to electronic health records has preferred expediency via PDF reporting and the like rather than maintaining structure through these transmissions. We propose the development of an SMO (site management organization) that has the ability to pull from both datasets and can therefore have a more structured view on large amounts of data in the healthcare system. This SMO would have the ability to link back to the longitudinal patient record to join genetic sequencing information to other clinical laboratory data on unique patient identifiers. For example, in the instance wherein a drug development company is looking for an individual with a high probability polygenic risk score for heart disease and a total cholesterol reading of greater than 250 in the last twelve months, then we will have positioned ourselves to provide an unparallel view. We can couple this with machine learning techniques to provide insight on “nearest neighbors” that would not be human intuitive in such a high dimensional data structure.
Coffee to be served outside FN 2.102 30 minutes prior to the talk.