Characteristics of health data: structure and relationships.

The characteristics of health data are diverse and vary by discipline. Health data can be viewed from three perspectives. They can be examined at a basic level of different variable types and database constructions. At an intermediate level, the way data are structured in a facility is a model of how an organization works. A computerized information system is a model of how work is performed. An understanding of facility data architecture and data structure is essential for adapting both manual and computerized systems to the rapid changes occurring in health care today. Lastly from the highest-level perspective, for communication and exchange of health data to be possible across all components of the health care delivery system, standards for data must be adopted by all institutions and individuals involved in the development of health data information systems. To expedite such exchange, the American Medical Informatics Association (AMIA) has recommended standards for patient, provider, and site-of-care identifiers; computerized message exchanges; and medical record content and structure. AMIA has also recommended the use of the universal medical language system as the basis for future developments in a universal medical language.