Children's Computing Resource Center

The PTTF recommends establishing an Children's Computing Resource Center (CCRC). The CCRC would facilitate coordination of data management and computing resources at CMH. Current, stand alone projects would benefit from integrated approaches or solutions. The benefits would be:
  1. Single source for data (personnel directory, patient diagnosis and medication lists, research protocols, etc.)
  2. Core software to support education and research
  3. Core personnel resources to support data analysis, biostatistics, data extraction, etc.
Organizational Structure

The current organizational structure cannot meet the information and computing needs of all constituents. As a consequence, many maverick solutions are being invented. A relatively simple example are the many database directories of people. There are about six different physician directories, each with different content. The time and resources to maintain these is large. Modern database technology would permit the use of a single, large database to which various constituencies contribute and withdraw data. The single source would be much easier to maintain while also being more complete and accurate.

The PTTF is recommending the creation of the Children's Computing Resource Center (CCRC). One basic function of the CCRC is to create a single source data repository for information that has multiple uses. The model requires constituencies to contribute and extract data using defined business rules. Each entity must be assured that it's rules are being followed by the other constituents. Such systems require controlled access in which users are authenticated and their access rights verified and restricted according to the rules. Because users may wear different hats (e.g., clinician and researcher), data must also be available for different purposes. For example, data for research may need to be de-identified before presentation to the user.

Technically, these arrangements are possible. The implementation requires very careful planning; creation of appropriate infrastructure, including administrative resources; regular exchanges of datasets; and continuous monitoring. The added value of data sharing justifies the effort. Traditional boundaries are blurred in such a system and a new organizational structure is required. Several models were considered, but the PTTF found most merit in the Hub and Spoke Model (figure).

Shared equipment and personnel in the CCRC would also produce economies of scale and create a Center that is greater than its parts.

David A Stumpf, MD, PhD

Informatics Program

Rationale for a Bioinformatics Core

Data Management Center.

Assuring the integrity of relational data requires rigorous attention to data tables, relational fields, and careful management of databases. Ideally, single entry is possible if applications can access the same data. Distributed storage is a leading edge technological innovation {see Techology Review (2004) 107:32}. Creating a dynamic CCRC would permit CMH faculty to lead multi-institutional research projects and manage the data

Data Analaysis. Researchers formally analyze their data using biostatistical methods. Standard and conventional software is readily available and runs on desktop computers. More sophistocated and complete analyses are not readily available and require specially trained personnel.

Data Mining. Analyzing extracted data from the Clinical Data Repository at Children's Memorial Hospital offers great potential for extending our knowledge about diseases in children. However, the commercial marketplace for vendors, such as Epic or GE, has not created sufficient demand for data extraction from data repositories. This has discouraged the entry of clinical data into CDR components as structured data elements which could be extracted for clinical research. Indeed, the GE PACS system, for example, has no capabilities for even sorting cases by diagnosis! Such diagnoses are embedded in radiologists' reports, and data mining is not possible with GE PACS products.

Specific Proposal for the Bioinformatics Core

A core bioinformatics facility would allow researchers at Children's Memorial Hospital to realize the potential for using clinical data from the CDR for quality improvement and research. We propose a bioinformatics core that will provide support for transfer of data from the CDR in a form that can be used for answering specific questions about disease and health. This core would have four components: technical; informatics epidemiology; pharmaceutical tracking; and decision analysis. This core would provide technical experts who can provide coding for data mining that would be placed on top of the clinical data repository. They would work in response to specific requests, to be approved by mechanisms to be established. In addition the core would provide staff expertise in epidemiologic evaluation of large data sets covering specific disease populations whose data will be populating the CDR. A third section of the core will provide integration of data from the CDR with tracking of progress of drug studies conducted by investigators at Children's Memorial Hospital. The final section will provide analysis of effects of data integration on decision support. This section will work in collaboration with the quality sections of the Hospital, and outside investigators, such as those at Northwestern's Institute for Health Services Research and Policy Studies (IHSRPS).

Significance

The establishment of a bioinformatics facility will establish Children's Memorial Hospital as a unique source of pediatric informatics research. This facility will directly serve Children's mission, which is to be dedicated to the health and well-being of all children, and includes being a leader in research into the prevention, causes and treatment of diseases that affect children. This facility will allow investigators at Children's to be even more productive in their research, in terms of grant support (such as that provided by the National Library of Medicine, in its extramural informatics program) and attracting additional talented faculty. The result of this facility will be enhancing CMH's contribution to improving children's health.
Michael Miller, MD; David A Stumpf, MD, PhD