Evaluation of a proposed method for representing drug terminology.

In the absence of a single, standard, multipurpose terminology for representing medications, the HL7 Vocabulary Technical Committee has sought to develop a model for such terms in a way that will provide a unified method for representing them and supporting interoperability among various terminology systems. We evaluated the preliminary model by obtaining terms, represented in our model, from three leading vendors of pharmacy system knowledge bases. A total of 2303 terms were obtained, and 3982 pair-wise comparisons were possible. We found that the components of the term descriptions matched 68-87% of the time and that the overall descriptions matched 53% of the time. The evaluation has identified a number of areas in the model where more rigorous definitions will be needed in order to improve the matching rate. This paper discusses the implications of these results.

The need for a concept-based medication vocabulary as an enabling infrastructure in health informatics.

Users of drug information typically focus their attention at different levels of description in different situations, such as medication ordering or dispensing. Computer systems utilizing drug information to support such activities must accommodate these multiple perspectives. This paper presents an approach to conceptualizing drug descriptions at multiple levels and outlines key features of an underlying information model that can serve as the basis for a concept-oriented medication vocabulary. These features include dose forms, routes of administration, as well as links to multiple drug classification schemes and medical problems. Implementation, standards, and maintenance issues related to the model are also discussed.

The future of knowledge-based components in the electronic health record.

Traditional monolithic healthcare information systems (HIS) no longer meet the requirements of today's distributed enterprises and the rapidly changing healthcare environment. The ability of applications to communicate, interpret, and act intelligently upon complex healthcare information has assumed paramount importance. The future lies in the development of flexible component-based architectures that can operate seamlessly within the workflow of a healthcare environment. A key design goal is "graceful degradation," i.e., providing the best decision support possible within the context of available patient data. The First DataBank Drug Toolkit is used as a case study. Several technical challenges associated with building truly plug and play components are discussed.

Towards improved knowledge sharing: assessment of the HL7 Reference Information Model to support medical logic module queries.

Because clinical databases vary in structure, access methods and vocabulary used to represent data, the Arden Syntax does not define a standard model for querying databases. Consequently, database queries are encoded in ad hoc ways and enclosed in "curly braces" in Medical Logic Modules (MLMs). However, the nonstandard representation of queries impairs sharing of MLMs, an impediment that has come to be known as the "curly braces problem." As a first step in solving this problem, we evaluated the proposed HL7 Reference Information Model (RIM) as a foundation for a standard query model for the Arden Syntax. Specifically, we analyzed the MLM knowledge base at the Columbia-Presbyterian Medical Center and compared the queries in these MLMs to the RIM. We studied 488 queries in 104 MLMs, identifying 674 total query data elements. Laboratory tests accounted for 45.8% of these elements, while demographic and ADT data accounted for 37.6%. Pharmacy orders accounted for 10.5%, medical problems for 4.3% and MLM output messages for 1.6%. We found that the RIM encompasses all but those data elements signifying MLM output (1.6% of the total). We conclude that the majority of queries in the CPMC knowledge base access a relatively small set of data elements and that the RIM encompasses these elements. We propose extensions of this analysis to continue construction of an Arden query model capable of solving the "curly braces problem."

Clinical decision support for physician order-entry: design challenges.

We report on a joint development effort between ALLTEL Information Services Health Care Division and IBM Worldwide Healthcare Industry to demonstrate concurrent clinical decision support using Arden Syntax at order-entry time. The goal of the partnership is to build a high performance CDS toolkit that may be easily customized for multiple health care enterprises. Our work uses and promotes open technologies and health care standards while building a generalizable interface to a legacy patient-care system and clinical database. This paper identifies four areas of design challenges and solutions unique to a concurrent order-entry environment: the clinical information model, the currency of the patient virtual chart, the granularity of event triggers and rule evaluation context, and performance.

The 'offset effect' of mental health treatment on ambulatory medical care utilization and charges. Month-by-month and grouped-month analyses of a five-year study.

The five-year ambulatory medical care experience of 400 patients with mental disorders was studied to test the "offset" hypothesis that patients receiving timely mental health specialist treatment have lower subsequent utilization of, and charges for, care than patients not receiving such specialist treatment. Specialist treatment was associated with significant offset savings in utilization and charges for the non-psychiatric medical care of treated patients with either severe or less severe mental disorders. However, the visits and charges for such specialist treatment boosted the overall (nonpsychiatric plus mental health specialist) care utilization and charges of the specialist-treated patients above those of patients treated solely by their nonpsychiatric physicians; this overall increase was especially pronounced for patients with severe mental disorders. The findings suggest the need for randomized prospective offset studies comparing utilization, cost, and clinical outcomes.