Biology of Blood and Marrow Transplantation | 2019
Design and Implementation of a Multipurpose Hematopoietic Stem Cell Information System Based on the Biomedical Research Integrated Domain Model
Abstract
Many hematopoietic stem cell programs capture transplant-related data into multiple data repositories simultaneously, including electronic medical record systems, repositories to support specific research projects, repositories to support specific quality management initiatives, repositories to track and guide the pre-transplant evaluation and workup, repositories to guide stem cell collection and harvest, repositories to support reporting to internal and external regulatory entities, and others. This state of affairs increases costs and decreases productivity as a result of marked duplication of data-capturing effort and poor quality, accessibility, and interoperability of captured data. An important barrier to addressing this issue is the fact that most data repositories in academic medical centers are based on different underlying data models, which constitutes a so-called chasm of semantic despair with regard to sharing and/or combining data residing in two or more data repositories, both within and between institutions. To address this issue, NCI, FDA, ISO, HL7, and C-DISC have developed a Biomedical Research Integrated Domain Group (BRIDG) Model. The purpose of the BRIDG model is to bridge the multiple chasms of semantic despair that exist between data repositories maintained by basic researchers, translational researchers, clinical researchers, pharmaceutical companies, and government regulators. To our knowledge, we are the first academic cancer center in the world to design and implement a cancer research system based on the NCI-BRIDG model. Important aspects of our implementation derived from the BRIDG model include extensive utilization of the NCI/BRIDG standard data elements, and resolution of the many-to-many relationship between patients and treatment protocols with a patient treatment course instance table. The latter feature provides powerful support for quality assurance of both clinical documentation and patient management by enabling our software to continually compare what is supposed to be happening to each patient, based on the treatment protocol or protocols on which the patient is registered, against what the clinical documentation indicates is happening to each patient. In addition to enhancing data sharing between researchers, we have also begun to use the system to support clinical decision-making, administrative decision-making, and clinical quality assurance.