The seamless integration of knowledge and data is indispensible to today's modern healthcare decision support systems. A healthcare organization that thoroughly understands its patients and is able to respond quickly to their needs, scores highly with them—and this has become an extremely important competitive component in today's ever-more interconnected world where patient feedback can positively or negatively affect an organization's reputation and bottom line.
The patient care world is complex, with various information systems being utilized to streamline and automate patient care processes. Fortunately, there is a new approach to IT efficiency vis-à-vis ontological engineering—or ontology programming—that is possibly the most significant benefit to ensuring accurate data integration, which fosters a better understanding of patient needs, thus resulting in better patient care and excellent patient outcomes.
Ontological engineering excel at extracting knowledge and critical information from the various information systems within a healthcare decision support system (or its organizational databases). Ontology programming reduces often difficult data integration issues and promotes data reuse, data sharing, and common vocabularies between the information systems, from patient intake to patient discharge.
For healthcare organizations to understand their patients better, data across the entire organization or spectrum of information systems involved in patient care must to be analyzed. Knowledge from different areas or "domains" (e.g., the patient-entry process domain, hospitalization and treatment domains, and billing and insurance domains) must to be extracted in order to accurately interpret quality of care.
Detailed knowledge is also required to interpret patient responses to the various care options exercised from the time of entry into the healthcare facility through final discharge. In addition, quality healthcare organizations strive to improve their existing processes and analyze post-care data in order to determine areas of improvement and initiate appropriate programs. Therefore, the accurate compilation and correlation of patient data is essential during the care process—both individually and in aggregate with other patient data—to determine potential process improvement steps.
As mentioned previously, healthcare organizations also benefit from their patients' recovering better and more quickly as a result of higher quality care. This is, in no small part, driven by efficient information systems. Patient care results are reflected in quality reports issued by premier organizations such as JCAHO (Joint Commission for Accreditation for Healthcare Organizations). As of 2009, JCAHO reports include patient satisfaction data, as well, thus making it even more important to understand patient information effectively and utilize to it to render care that leads to better patient satisfaction.
Accurate knowledge across intra-organizational domains can only be extracted when healthcare decision support systems are able to exchange relevant data with each other—which is not always possible with current configurations. Even if the numerous systems within an organization can connect to each other through common computer interfaces, they may have stored patient data differently, rendering information exchange virtually impossible and creating a silo effect. Additionally, the context in which the information is used may vary from system to system, making it even more difficult to correlate data across various platforms and systems within the organization. Finally, data consistency and data integrity issues arise as each silo information system is further customized to optimize the information system's performance.
Therefore, to achieve a comprehensive and accurate individual patient view across the entire patient care spectrum of an organization, different information systems-based reports may have to be compiled separately with data correlated between them. The results will then need to be represented in a single, coherent report. This type of data correlation may include the mapping of various customer names for a single patient, as an example. Obviously, this type of system is not only vulnerable to error and to data integrity and consistency issues, but it is also quite inefficient and, therefore, needlessly costly.
Data correlation, integrity, and integration issues are not confined within an organization's systems only. Health care organizations rely on HIE (Healthcare Information Exchange) to communicate with external entities. HIE is used to move clinical information between different information systems from various providers (i.e. test labs, insurance companies, and other healthcare facilities) without losing the meaning of the information exchanged. These systems typically use established standards for data exchange, such as SNOMED CT, ICD-9 and ICD-10, and other HIE standards.
Periodic updates are required, and organizations must ensure that they are in compliance in order to participate in data exchanges with other providers. Naturally, whenever any data changes occur, the cost and time required to modify multiple systems within an organization can be staggering, but without the use of ontological engineering, the higher costs must be borne, as system modifications are mandatory.
Whether the data reside internally or external sources are employed for HIE, a healthcare organization faces the common issues of data mapping, data integration, reuse, and data sharing. Whenever data change, or new relationships between data are discovered, organizations expend valuable resources in time and money adjusting databases across various systems in an attempt to keep them aligned with each other. This absorbs important resources, taking them away from the core focus and value proposition of the organization—that of providing quality patient care.
When data change, especially internal organizational data, conventional technologies (as in "relational" databases) require changes to their database structures and schemas, potentially leading to major regression testing of the systems after the changes have been completed. This must be accomplished in order to ensure that nothing is deleted or corrupted after the changes are made, and is quite naturally, another costly step—both in terms of time and resources.
Information technology departments have tried to respond to data integrity and data integration issues across various systems within an organization by building a data warehouse that acts as a central repository for most, or all, of the inter-related systems. However, the solution is only partially successful. Often times, competing interests from various internal "stakeholders" in different information systems can lead to data that is stored in a manner is favorable to some information systems, but not others. This, of course, potentially compromises data access and reuse by other systems.
In addition, since the entire organization's data cannot be migrated to a data warehouse simultaneously, some systems are migrated before others, and the entire migration process may take as long as a year or more to complete in a large health care organization. In the interim, data across the enterprise changes, and the whole cycle of re-aligning data must start anew. There have been proposed solutions to address this and other related problems, but they each leave something to be desired.
Ontology programming can help reduce data integration, sharing, and reuse pains to quite an extent. By definition, they are "a formal representation of knowledge by a set of concepts within a domain." They not only store data in a database, but also store relationships, including hierarchical relationships, between data.
This ability distinguishes ontological software from standard relational databases and provides the flexibility of updating data and relationships between them. Ontologies are also able to add newly discovered relationships without the necessity of significantly changing the core database or requiring extensive programming efforts—unlike typical databases currently in use. They also excel at removing term confusion and providing data mapping capabilities, which vastly promotes improved data share and data reuse across an organization's information systems.
For healthcare organizations, as well as other large business enterprises, the practical, time-saving applications of a system built on ontology programming are quite extensive. We know that ontological engineering provides the ability to extract knowledge contained within applications and information systems across the various domains within an organization, but it is also very useful for capturing "real world decisions" made by humans and converting it into computer format. The result of this capturing of knowledge across domains by SMEs (Subject Matter Experts) and healthcare providers leads to much more consistent query results whenever similar conditions are encountered in the future.
Such information system architecture can significantly reduce medical errors and enhance patient care. This can be accomplished, for instance, by the capturing of a healthcare professional's diagnosis of a particular medical condition and other relevant data. Once the data are entered into the ontological system, it will consistently provide the same results for similar conditions in the future and offer the diagnostics and conclusions as an aid to other healthcare professionals.
Subsequently, a healthcare professional may choose to exercise the same diagnostics (or treat the patient differently according to differences in patient circumstances), but the healthcare decision support system's information can now provide an important, relevant checkpoint based upon the previous diagnostic information.
In conclusion, the use of ontology programming in the healthcare field provides a significant reduction in data integration issues and—because semantic technologies are superior extractors of knowledge across multiple information systems and can add new relationships between such systems with relative ease—they provide the flexibility to change data with far less effort and cost than standard systems now require.
Consequently, ontological engineering are able to provide an invaluable component to improved patient care and outcomes by supporting critical healthcare processes and decision-making. The superior integration of knowledge and data within healthcare organizations may at first appear prosaic, but it is nothing short of revolutionary in its potential to affect organizational performance and quality care.