3 Note that the two-character codes used in Internet URLs (such as the "uk" in the URL of web site of the government of the United Kingdom: are ISO 3166-1 two-character country codes.

4 For example, in 1989 Burma changed its name to Myanmar and the ISO 3166-1 two-character code changed from BU to MM.

5 For example, Czechloslovakia (former ISO 3166-1 code CS) split in 1993 into the Czech Republic (ISO 3166-1 code CZ) and Slovakia (ISO 3166-1 code SK).

builds clinical information systems has implemented its understanding of the standard in its products.

Health Level Seven, Inc. created the first version of HL7 in 1987. Since then, it has continued to develop and maintain the HL7 message standard. Health Level Seven, Inc., became an American National Standards Institute (ANSI)-accredited standards-developing organization in 1994. It has 26 international affiliates, meaning that organizations in 26 countries are using the standard and contributing to its ongoing development. The American Veterinary Medical Association endorses HL7 for use in veterinary information systems. HL7 is not free; one purchases documents that specify the standards (members of HL7 receive a discount).

HL7 is event based, meaning that as soon as an information system completes a task, it generates and transmits an HL7 message to other information systems. For example, after a registration clerk registers a patient in the emergency department (ED), the registration computer transmits a message to other systems such as the laboratory and radiology information systems (to prepare them to receive orders or specimens for the patient). Figure 32.1 shows a typical de-identified ED registration message that hospitals send to the RODS biosurveillance system. Similarly, laboratory and radiology information systems transmit messages about test results as soon as the results become available. HL7 defines as "trigger events'' various tasks or transactions, including patient registration, orders, and results. HL7 defines nearly 90 types of messages. For each type of trigger event, HL7 specifies which data elements should be in a message, where they may be located in the message, and whether they are required versus optional.

Health Level Seven, Inc., began work on version 3 of HL7 in the late 1990s to remove the optionality present in the 2.x versions of HL7. The goal of version 3 is to improve interoperability of clinical information systems. Version 3 HL7 also includes a semantic model called the HL7 Reference Information Model or RIM (see below). HL7 has formally approved and released portions of the version 3 standard, including the HL7 RIM.

If you are building a biosurveillance system that connects to hospitals, it is a certainty that you will encounter HL7. It is the most widely implemented language standard in health care.

As with SNOMED-CT and LOINC, however, it will be several years before you encounter healthcare organizations that are capable of transmitting data by using version 3.0 of HL7. Vendors only recently have begun implementing HL7 version 3.0 in new versions of their products, and healthcare organizations still must then upgrade to those new versions and use version 3.0 of HL7 in place of version 2.x.

4.2.2. Accredited Standards Committee X12 Insurance Claims Standards

Standards of the Accredited Standards Committee X12 (ASC X12) are message format standards for information systems in a number of industries. The X12N subcommittee of ASC X12 creates standards for insurance transactions. ASC X12 standards are not free; one must purchase the documents that specify the standards.

We mention ASC X12 here because the Department of Health and Human Services has mandated ASC X12 standards for the submission of healthcare insurance claims as part of the administrative simplification provisions of the Health Insurance Portability and Accountability Act. Developers of biosurveillance systems may thus encounter ASC X12 when using claims data from physician offices or hospitals for surveillance.

ASC X12 has signed memoranda with Health Level Seven, Inc., and the National Council on Prescription Drug Programs (see below) to ensure that its standards do not overlap and/or are compatible with the standards of those organizations.

4.2.3. Pharmacy Claims Standards

The standards of the National Council for Prescription Drug Programs (NCPDP) are message format standards for transmitting data as part of commercial pharmacy operations. The Batch Transaction and Telecommunications standards are message formats for submitting claims to healthcare insurance

MSHiA--\&j|ixx||RODS|2Q0202241715| ADTAA04i2Ö02022XXXXXXXX!P}2.3<C R>

PVl||E||||l||||||||||||98765432|||||]||||||l||[||||||||200202XXXXXXXi|<CR> DG1IIICARBON MONOXIDE EXPOSURE<CR>

INI|!lll|||||!l||[ll|||||ll|||lll!l||ll|||Hlir^84056<CR> <ETX>

figure 32.1 A typical HL7 message for transmitting chief complaint data to the RODS biosurveillance system. The message contains date of service (February 24,2002 at 17:15 hours), message type (ADT 04 standing for emergency department registration), patient age (20), gender (M), chief complaint (CARBON MONOXIDE EXPOSURE), and home zip code of patient (84056).

companies on behalf of individuals with prescription-drug coverage. The SCRIPT standard is a message format standard for transmitting electronically prescriptions from a physician's office to a pharmacy.

The NCPDP has created and maintained pharmacy transaction standards since 1977. Similar to HL7 standards, NCPDP standards are not free. You must either be a member of NCPDP or purchase the documents (which come with a membership to NCPDP) that specify the standards.

You may encounter NCPDP standards if you are using pharmacy claims data for biosurveillance because the Department of Health and Human Services mandated the Batch and Telecommunication Standards for use with pharmacy claims submission as part of regulations it developed under the Health Insurance Portability and Accountability Act.

4.2.4. METAR Meteorological Observations Message Standards

METAR refers to the METAR message-format standard—as well as its vocabulary and semantic standard—for hourly observations of surface weather. It specifies a strict ordering of the data elements in the message, as well as which data elements must appear and which are optional. For example, the weather station code is mandatory and always appears first in the message. See the discussion of METAR under "Standard Vocabularies'' for more details about METAR and its role in biosurveillance.

4.3. Semantics Standards

Semantic standards specify additional rules for constructing messages. In human languages such as English, semantics refers to a set of rules that humans intuitively understand, and therefore (unless you are a linguist), you will not have encountered any semantic rules in your life (except indirectly when you learned English as a very young child or as a second language). The following sentence adheres to English grammar and vocabulary rules but violates semantic rules: Blue mathematics eat televisions. A speaker of English realizes that the verb eat cannot have blue mathematics as a subject or a television as its direct object.

Because computers do not have such intuitions, language standards must specify that when the verb is eat, only a limited number of subjects and direct objects are legitimate. In the clinical domain of laboratory testing, for example, you must tell the computer that laboratories "test'' specimens from patients. Only laboratories can "test,'' and the direct object of the "testing'' act is a specimen. You must then also tell the computer which vocabulary codes are for laboratories, which are for patients, and which refer to specimens. This latter function is accomplished by the vocabularies, but the relationship between the action verb test and the legitimate classes of entities is the job of semantics.

A standard information (or data) model is a semantic standard. It specifies for a domain the set of verbs or acts, the set of allowable actors for each act, and the set of allowable direct objects for each act. It also specifies the other data elements (analogous to adjectives, adverbs, and prepositional phrases) that qualify each act, such as when (date/time) and where (location) it occurred.

The HL7 RIM is a standard information model (semantic standard) for medicine. It specifies the acts, actors, and entities acted upon and the data elements associated with each that are necessary for clinical information systems to communicate data. The significance of the HL7 RIM is that it is the key to eliminating ambiguity in HL7 caused by the lack of semantic standards; thereby eliminating a source of incompatibility between systems. Health Level Seven, Inc., approved and released the HL7 RIM in 2003.

Because HL7 RIM is the core of version 3.0 HL7 standards, it is necessary to understand the RIM to create and to process HL7 version 3.0 messages. However, as we discussed previously, it will likely be some time before you encounter clinical information systems capable of sending version 3.0 messages.

4.3.2. Public Health Information Network Logical Data Model

The Public Health Information Network (PHIN) Logical Data Model (LDM) is a standard information model of public health. It is an extension of the HL7 RIM to accommodate public health "acts'' such as case detection and outbreak detection. The Centers for Disease Control and Prevention (CDC) released version 1.0 of the PHIN LDM in 2004.

The PHIN LDM is still in relatively early stages of development. Because the development of standard data models is as much an art as it is a science (and moreover is a highly iterative process), until developers use the PHIN LDM in biosurveillance applications, the validity of the modeling decisions and their ability to facilitate data sharing will not be known.

Nevertheless, the PHIN LDM may serve as a useful starting point in creating a database to store biosurveillance data (provided that the PHIN LDM covers the acts, entities, roles, and data elements you need to store the data you are collecting). Wilson et al. (2001) describe in detail how they derived a LDM from a predecessor of the PHIN LDM.

4.3.3. METAR Semantics

METAR is the semantic standard for hourly observations of surface weather (as well as the vocabulary and message format standard). It specifies various rules for constructing messages about weather observations. For example, it specifies that messages about observations made at a particular time of day must include certain data elements. For example, messages about observations made during the 12:00 UTC hour must contain the amount of rainfall that occurred in the preceding 24 hours. See the discussion of METAR under "Standard Vocabularies'' for more details about METAR and its role in biosurveillance.

Was this article helpful?

0 0

Post a comment