Biomedicine is currently undergoing an unprecedented transformation from an observational science to an informational science. This transformation is marked by a variety of changes, such as:

• Rapidly Increasing Adoption of the Electronic Health Record (EHR)
• • The EHR was quickly adopted in the healthcare community and the average amount of electronic data archived per person is rapidly increasing.
  • Increased use of the EHR applies to all aspects of healthcare. Patient data is interactively captured and accessed in real-time at the point-of-care in the following forums:
  • • Emergency room visits
    • Inpatient stays
    • Outpatient visits
    • Primary care services
    • Routine office visits
    • Pharmacy requests

• Lifetime-Based EHR Capture
• • As a consequence of Hospital-wide adoption of the EHR, individual patients are treated based on lifetime health records, rather than their most acute health events.
  • This lifetime archive provides information that can lead to better diagnosis and treatment for patients.
  • The lifetime archive is also a rich resource of data for research into behavioral, environmental, and genetic causes or treatments of disease.

• Increase of Data-Intensive Technologies
• • In addition to the surge of data from an increasing number of EHR-captured patients and the lifetime of data collected for each patient, this rapid increase in data is also caused by the advent of multiple data-rich technologies, which include various imaging technologies, whole-genome assays such as arrays and mass nucleic acid sequencing, and molecular structure-based assays.
  • These technologies are critical to improved diagnosis and treatment of disease and are promising vehicles for future biomedical developments; however, they add huge data management challenges in terms of the amount of data captured per assay.

• Complexity of Biomedical Data Integration
• • The types of data captured for patients are highly variable in terms of size, procedures of storage and extraction, and degree of support required for proper interpretation of results.
  • Take, for example, the difference between the requirements for interpretation of a simple blood pressure assay and those of a gene expression microarray-based result.
  • • Blood Pressure:
    • • Simple numerical data
      • Quickly stored and accessed
      • Requires no additional support for interpretation
    • Microarray (i.e., hundreds of thousands of interdependent measurements):
    • • Data is stored in any number of formats including images (gigabytes/terabytes) or a matrix of pre-normalized signal values (megabytes)
      • Requires specific procedures for storage and extraction
      • Only interpretable by a highly specialized analyst or a pre-fabricated application
  • The consequences of having these two types of data, and hundreds of other types, yield highly complex database designs, access procedures and interpretation mechanisms.
  • Public Health Policy and Initiatives
  • • In addition to the technical challenges associated with this transformation, public health policymakers are striving toward appropriate management of such issues as data accessibility, data usage and privacy concerns.

  At The Children's Hospital of Philadelphia, there is abundant evidence that the data portion of the health informatics transformation is already occurring. In research labs, technicians are more often at computer terminals than at the lab bench. Similarly, clinical space, workflows, and activities revolve around computer terminals and computational representations of patient information. Even patients and families are turning increasingly to computer tools to understand health and disease processes. A variety of Hospital resources, such as the Stokes Microarray Core Facility, the imaging facilities in the Department of Radiology, and the clinical instrumentation services for monitoring inpatient vital signs, are already generating large and complex data sets. These trends will continue as data generating capabilities continue to evolve.