Development and Utility of a Novel Intergenerational Health Knowledgebase

Abstract

ObjectiveOur long‐term goal is the development of a clinically‐relevant, validated enterprise data warehouse for research. The Intergenerational Health Knowledgebase (IHK) will advance and accelerate the study of the developmental origins of adult disease. We sought to harmonize data from the electronic health record (EHR) and external data sources and efficiently synthesize key variables from pregnancy and early life in the IHK. The objective was to make the IHK be efficient, effective, flexible and able to integrate with participation information for on‐campus biobanks.MethodsThe development of the IHK is an iterative process involving multiple steps: 1) clinical data team (CDT) members identify and locate clinically significant distinct clinical variables in the front‐facing EHR pertaining to obstetrics and pediatrics. 2) The Biomedical Informatics group in the Institute for Clinical and Translational Science (ICTS‐BMI) group locates these variables in the transactional database of the EHR. 3) The CDT members validate the extracted variables. Extracted data is compared with expected outcomes from the EHR. Inconsistencies are reviewed by the CDT and ICTS‐BMI group to pinpoint potential causes. 4) With confirmed data variables, ICTS‐BMI develops novel data tables and user‐friendly SQL queries. The resultant OBstetrics Data Integration Architecture structure integrates multiple data sources within the institution which includes, but is not limited to, imaging data, diagnoses, vital sign information, medications, medication administration, and procedures for the maternal‐fetal dyad. Powering this architecture is a common semantic layer that uses established standardized vocabularies and value sets. This includes storing patient demographic data, patient diagnoses using ICD9‐10 codes, storing laboratory results using LOINC codes, and problems using ICD9‐10/SNOMED‐CT codes. In our EHR, all data is joined using pregnancy episode identification numbers linking maternal and neonatal data under a singular pregnancy instance. ICTS‐BMI developed a secure web‐based Knowledgebase viewer for performing queries of the IHK without expert SQL knowledge.ResultsThe IHK contains indexed maternal and child data. The IHK is frequently updated to contain every pregnancy cared for at the University of Iowa Hospitals & Clinics (UIHC) (currently 62,000+ pregnancies). Because of our team’s complementary expertise, they were able to anticipate needed information, identify and validate key variables in the EMR, and plan pre‐defined queries. As a result, we have quickly provided data for studies ranging from the effects of maternal environmental exposures, the role of patient distance to the hospital in pregnancy outcomes, medication use during pregnancy, and the association of depression with adverse outcomes.ConclusionThe development of the novel IHK has facilitated pregnancy‐related research and identified new targets for improving maternal health and following long‐term health outcomes of parents and children. Further, integration of the IHK with the UIHC Women’s Health Tissue Repository facilitates using biological specimens for research.