Knowledge base technology for healthcare: From True or False to True

By the 1950s, computer science had embraced Boolean logic, named for the nineteenth-century mathematician, George Boole. His logic reduces all values to TRUE or FALSE, which meshes with the binary numbering system used by computers in which each bit is valued as either 0 or 1.

In 1974, IBM’s research centre evolved Boolean logic into a concept that emerged as the structured query language, SQL, now widely used to access records in databases in all sizes of computers. Accessing the records is speeded up by indexing performed by the database management system, which also understands SQL.

In the early 1990s, the World Wide Web surfaced as new technology for accessing information. As a system of internet servers offering documents formatted in hypertext markup language—a derivative of an international system for document printing—the Web spawned search engines. These examine the documents for specified words or phrases and then fetch copies of documents that contain them.

The search engine technology indexes the documents so that those presented to the human searcher are as useful as the computer can make them. Thus the fetch processes of the Web, especially the all-important indexing, rely on computers.

Paralleling the Web’s extraordinary growth was the equally dramatic uptake of the graphical user interface, used in the Macintosh and Windows operating systems. Less spectacular but no less fundamental was the popularization of the treeview, an arcane computer construct enabled by the graphical user interface. Now familiar to millions and millions of people working with personal computers in so many of the world’s human languages, the treeview—a filter system— underpins the quiet revolution in information science.

The quiet revolution began in the late 1990s with the burgeoning requirement for precision: on the search engines’ fetching was imposed the need for more and better filtering. From that need grew the computer industry’s construct of the knowledge base.

Knowledge base is a term widely used in non-computer contexts. In the computer world it increasingly refers to (a) a body of knowledge including knowledge in people's heads and their understanding of that knowledge, (b) a particular browser format, and (c) a treeview approach to the organization of the knowledge in which indexing is primarily in the hands of subject-area experts and not left entirely to the computer.

For healthcare, the three computer methodologies differentiate into three largely distinct usages.

The database, with its capacity for precision in locating individual records, finds its most promising application in storing private and personal and medical data—such as lab results and lists of medications prescribed—relating to individual patients. Modern database technology reduces the probability of attribution of your data to someone else, and of the wrong person’s viewing your personal data.

Search engines used with large databases of the published medical literature as well as the Web most usefully serve researchers by fetching studies containing particular words or phrases. Search engines offer no assurance that the fetch includes every study that is relevant, especially because relevance is a characteristic that the only the researcher can fully judge. Filtering and refining the yield, time-consuming and complex, is a skill honed by researchers for whom fetch-oriented technology of search engines is best suited.

Meeting the knowledge needs of practicing healthcare professionals is the task for the knowledge-base approach. Recognizing that health professionals can no longer be expected to carry in their heads everything they need to know about the healthcare they practice, the UK established its National Electronic Library for Health. While this is still mostly search-engine oriented, it’s fostering recognition that frontline professionals need the filtering and precision of the knowledge base.

On the healthcare front-line, professionals need the speed, convenience, and certainty of the point-click-expand-select of the treeview. To respect the filtrate, they need to see that the treeview that yields it is designed by professionals and subject area experts from their own fields. To trust it, they require it to unambiguously pinpoint the knowledge they can have confidence in. To press it immediately into front-line service, they need it be relevant and comprehensible.

Meeting front-line professionals’ requirements is the healthcare application of choice for the knowledge base.