As discussed previously, the standards for which you need passing familiarity include standards for data communication, networking, and knowledge.
5.1. Standards for Transferring Data to Data Analytic Programs
There are several standards for transferring data from a database to a data analytic program such as SAS and Crystal Reports or to geographical information system (GIS) packages such as ESRI's ArcGIS software.
5.1.1. Open Data Base Connectivity/Java Data Base Connectivity
Open Data Base Connectivity (ODBC) and Java Data Base Connectivity (JDBC) are standards that enable analytic (and any other) programs to retrieve data from relational databases. In general, each relational database vendor creates an ODBC and/or JDBC "adaptor'' for its database product. Programs compliant with these standards can access databases that are also compliant.
The relevance of these standards to biosurveillance is that many epidemiologists and other public-health personnel use data-analytic programs such as SAS and Crystal Reports to analyze biosurveillance data. If your biosurveillance database is ODBC compliant, users can access data from these programs. One note of caution is that some data requests that users submit via these programs may be so resource intensive to execute in terms of the processor, memory, and/or disk that they severely decrease the performance of a production biosurveillance system, hindering access and use by other users.
ArcGIS file formats are standard file formats that GIS programs use to store geospatial data. Geospatial data are often used to draw maps. Two important standard file formats that GIS systems typically have the ability to read and write include ESRI ArcImport and Shape files. Viewing biosurveillance data in GIS programs, such as ESRI's ArcGIS package, requires that the biosurveillance system format geographical data in a standard format understood by the GIS program.
Networking standards are standards for establishing a communication channel between computers. Networking standards are sufficiently mature and widely deployed that you only need to know that the hardware and software you use to create networks will be using standard protocols such as transmission control protocol/internet protocol (TCP/IP). One networking standard with which you may need to be familiar is virtual private networking (VPN), because you will have to make a conscious decision whether to use it and you will likely need special hardware and software to use it.
VPN is a standard method for establishing secure communications among applications over the Internet. The advantage of VPN is that it does not require any customized software additions to applications to enable them to encrypt data or to implement Internet protocols such as hypertext transfer protocol (http). VPN does this. VPN is useful, for example, when connecting a hospital's HL7 message router to a biosurveillance system located outside the hospitals' computer network. No modifications need to be made to the message router or its software to send data over a VPN.
You can choose between using VPN software or using dedicated VPN hardware. In our experience, dedicated VPN hardware provides a more reliable connection than does using VPN software on devices that also perform other non-VPN functions (such as a firewall). Many hospitals already operate a VPN device, making VPN the ideal method for connecting to such hospitals.
We discuss other methods for secure transmission of data over the Internet in Chapter 33 because the discussion is more relevant to the architecture of a system than to standards.
There are two types of knowledge standards relevant to biosurveillance: standard knowledge bases and standard knowledge representations. A standard knowledge base is an agreed upon set of facts or knowledge about the world represented in a way that a computer use it. A standard knowledge representation is an agreed upon way of representing facts and knowledge about the world in a knowledge base. Standard knowledge bases facilitate your work because you do not have to recreate the knowledge; you can just use it. A standard knowledge representation facilitates sharing of knowledge among knowledge bases. None of the standard knowledge bases we discuss here use a standard knowledge representation, thus the remainder of this section discusses knowledge bases that are either defacto standards or could become so through more widespread use.
The Notifiable Condition Mapping Tables (formerly the "Dwyer tables'') are a standard set of facts about which tests and results indicate that a patient has a notifiable disease. The Notifiable Condition Mapping Tables facilitate implementation of ELR by providing the "diagnostic'' knowledge that laboratory information systems need when determining which results to send to a biosurveillance system. For example, one table maps from LOINC codes to notifiable diseases. It maps LOINC code 22863-5—a test for antibodies to B. anthracis in serum—to the notifiable disease anthrax. Another table maps from SNOMED-CT codes to notifiable diseases. For example, it maps the SNOMED-CT code 21927003—the B. anthracis bacterium—to anthrax. The idea is that if the 22863-5 test is positive or a microbiology culture grows 21927003, then the laboratory information system should transmit the test and its results to public health.
As notifiable diseases vary from state to state, a laboratory would utilize the knowledge about only those diseases on a particular state's notifiable disease list. Specifically, the laboratory can pull the LOINC and SNOMED-CT codes that are associated with each disease that they must report. The laboratory can then filter its test results for the appropriate laboratory test data about notifiable diseases. These tables do not require a license fee for use.
A standard ontology is a standard specification of concepts and the relationships between concepts. The relationships between concepts encode knowledge, and thus, an ontology is a knowledge base. Many of the standard vocabularies we have discussed in this chapter are also ontologies because they provide relationships between concepts as well as concepts. Specifically, LOINC, SNOMED-CT, ICD, the NDC directory, RxNorm, and the UMLS all contain relationships between concepts and thus are ontologies. For example, SNOMED-CT encodes the following knowledge: the disease respiratory anthrax (11389007) has an "associated etiology'' (116675007) relationship with B. anthracis (21927003), meaning that B. anthracis causes respiratory anthrax. The code 116675007 uniquely identifies the "associated etiology'' relationship of SNOMED-CT.
The most common relationship between concepts in ontologies is the "is a'' relationship, which indicates that one concept is a type of another concept. For example, SNOMED-CT specifies that the disease respiratory anthrax (11389007) "is a'' (116680003) anthrax (17540007), meaning that respiratory anthrax is a type of anthrax.
This type of knowledge can facilitate creation and maintenance of the logical rules that a system follows to perform certain tasks. For example, if an ELR system needs to report all cultures that grow an organism of the genus Salmonella, then it is easier to write a rule about the SNOMED-CT code for Salmonella—27268008—and have the system check each organism code to see if it is a type of Salmonella than to write a rule with more than 100 codes for each species and strain of Salmonella. This approach also facilitates system maintenance: it requires no modification to the logic of your system when a new strain of Salmonella is added to SNOMED-CT.
With respect to biosurveillance, the utility of ontological knowledge about diseases, symptoms, drugs, and organisms depends on the application. At a minimum, when inferences about class membership (one thing being a type of another) are important, a standard ontology can greatly facilitate your work.
Was this article helpful?