Databases - Bioinformatics Computing - page 48

contribute to abnormal behavior. These and similar clinical data are stored in an electronic medical
record (EMR) in the format described in Table 2-2 .

Table 2-2. Typical Electronic Medical Record (EMR) Contents. The EMR
contains both objective signs, such as physical examination findings, as
well as subjective patient symptoms, including chief complaint and review
of systems.

Data Category Description

Chief Complaint Patient's primary reason for the medical visit

History of Present Illness History of onset of clinical signs and symptoms

Medications Current list of medications the patient is using

Past Medical History Relevant past medical history, including hospital admissions, surgeries,

and diagnoses

Family History History of family diseases, such as diabetes, cancer, heart disease, and

mental illness

Social History Use of drugs, smoking, job stability, housing, living conditions,

incarceration

Review of Systems Patient's recollection of symptoms and current medical problems, such as

trouble sleeping at night or panic episodes, and results of tests

Physical Examination The clinician's hands-on examination of the patient, including head, eyes,

ears, nose, throat, chest, and extremities

Labs Includes blood glucose, cholesterol, and drug levels

Studies X-ray, MRI, CT, and EKG

Progress notes Record of temporal progression of signs and symptoms, labs, and studies

for the length of the study or admission

The components of the EMR report rarely exist in a single, unified database, but reside in the
separate, domain-specific databases that may exist within a single hospital or clinic or be dispersed
geographically across a region or country. Regardless of their relative proximity to each other,
laboratory, radiology, cardiology, hematology, internal medicine, and other clinical departments
typically maintain their own medical-record systems. What's more, each application may be
supported by a different operating system, use a different underlying database—some of which may
be outdated—and execute on a completely different hardware platform. For example, the pharmacy
system might run under UNIX on a Sun Server using a Sybase database, whereas the clinical
radiology system might run under VMS on a VAX server with an Oracle database. Within each
department or clinic, these differences are usually irrelevant unless data have to be shared with other
departments. The traditional method of creating a composite view of a patient's clinical status is to
generate custom reports, which is time-consuming and expensive. The modern approach to the EMR
is to create one or more central databases derived from, and yet completely independent of, each of
the application databases, and to optimize these databases for research and analysis.

In order to create a comprehensive record that can be queried, the data from the various clinical
systems have to be integrated, usually with the assistance of a data dictionary that translates various
clinical databases to common formats so that the data can be more easily combined. The data
dictionary is, in simplest terms, a collection of information about naming, classification, structure,
usage, and administration of data that originates from a variety of sources. The data dictionary is
perhaps most useful in addressing the problem of data element ambiguity. For example, within a