Clinical Semantics - Computational Intelligence in Healthcare: Advanced Methodologies - page 9

As stated earlier, there have been considerable energies expended in an attempt to
standardize and formalize clinical information by developing encoding standards and
ontologies also for use in structured data entry and management (e.g., within the
Health Level 7 (HL7) and Unified Medical Language System (UMLS) initiatives).
However, some recent discussion in medical informatics has brought attention to the
fact that the overall attempt towards an efficient EMR, wherein this approach is an
integral component, has not proven to be a panacea. [9]
Information systems efficiently integrating with clinical work remain an elusive
goal. As Wears and Berg state in [13],
“Behind the cheers and high hopes that dominate conference proceedings, vendor
information, and large parts of the scientific literature, the reality is that systems that
are in use in multiple locations, that have satisfied users, and that effectively and
efficiently contribute to the quality and safety of care are few and far between”
In relation to this, EMR's must intertwine information from a multitude of domains
and levels of abstraction into a complex and evolving operational context. Clinical
operations and decision making requires the application of data that do not always
easily fit into a predetermined, rigid and linear set of selections. Restricted expression
of clinical information via formulaic constructs often results in a loss of expressivity
and limits the capture of important, subtle clinical details. A small fragment of natural
language can convey a complex message and is natively understandable, by defini-
tion, to humans. In addition, an ever-increasing number of complex user interface
components, as typically necessary for structured data entry in clinical systems, have
a negative effect on both end-user usability and data entry efficiency. Still, these
shortcomings do not take away from the fact structured data entry does suite a number
of EMR scenarios, just that it is not applicable universally. [9]
To further illustrate the challenge of clinical informatics, we will next look into
two clinical process transitions involving the use of structured electronic medical
records. The first one describes a case where a manual data gathering and manage-
ment process is replaced with a computerized one. The second one involves the archi-
tectural migration of an existing computerized clinical process.
2.1 Transition from a Manual to a Computerized Clinical Process
An evaluation study on changes in the quality of care and cost of care resulting from
the reorganization (supported by computerization) of a Finnish primary care antico-
agulation treatment workflow is described in [14]. In Finland, the anticoagulant effect
of warfarin is usually followed up with laboratory testing of venous blood samples
guaranteeing that the correct anticoagulation level, measured in P-INR units, is main-
tained. Follow-up is organized by public primary care health clinics covering the
anticoagulated population of a municipality or a number of municipalities [15]. Con-
sidering the follow-up media, data have typically been recorded on paper-based anti-
coagulation notebooks. However, primary health care EMR's increasingly offer basic
functionality for the maintenance of anticoagulation follow-up data. The study in
question involved the development of a separate, structured system for managing
anticoagulation data in an electronic format. Figure 2 shows the architecture of the
system.