|
Quality medical care in the modern hospital is
impossible without complete and reliable communication among the
constellation of caregivers who contribute to the treatment of each
patient. Medical advances cannot occur without consistent and comprehensive
methods for linking and comparing the experiences of many patients
across a spectrum of treatment settings. For both individual patient
care and the accumulated experience of medical knowledge, automated
data systems drive the transfer of information that improves treatment
and saves lives.
Yet many of the systems designed to streamline
data collection and process that information for retrieval and comparison
are unable to communicate with similar systems in other hospitals,
clinics, and research institutions. Some systems are unable to communicate
even within their own institutions. The problem? These systems lack
a common language and a unified set of criteria by which information
is labeled, accessed, stored, and retrieved. The result? Valuable
data are not communicated or, worse, are miscommunicated. Entire
databases on topics of vital clinical and research interest are
inaccessible or unintelligible to colleagues working on similar
problems.
In April 2002, the Anesthesia Patient Safety Foundation
(APSF; Pittsburgh, PA) and leading anesthesia information systems
manufacturers announced the launch of a coordinated effort to define
and address the problem of consistency and communication in surgical
anesthesia. The Data Dictionary Task Force (DDTF) will focus on
improving patient safety and medical communication by defining a
common set of data elements required in the electronic anesthesia
record. "A variety of computer-based anesthesia records is in use,
but they lack the ability to interface with each other or with a
central clinical data repository," says Terri G. Monk, MD, Professor
of Anesthesiology at the University of Florida (Gainesville) and
head of the DDTF. "The retrieval of information from these systems
is also limited by inconsistencies in the naming of problems, medications,
and other data."
The members of the DDTF are working to address
two key questions: What are the outcomes in the anesthesia setting
that should be investigated to yield beneficial information? And
what are the most important questions that should be asked and answered
through data collection and analysis to achieve the greatest immediate
benefit to patients? The answers to these questions will provide
the basic organizational structure on which a coherent data dictionary
for anesthesia can be built.
A lack of standards for the coding and exchange
of information plagues every medical discipline and treatment setting
and has been identified as a major source of medical error. The
members of the DDTF from throughout the anesthesia community hope
that by addressing this problem proactively within their own field
they will provide a useful template on which other disciplines can
base similar initiatives. "A common data set allows for the collection
and comparison of large volumes of clinical data from multiple institutions
for outcomes research and benchmarking," says Robert K. Stoelting,
MD, APSF president. "We are confident that this innovative joint
effort by our expert clinician panel and the manufacturers of anesthesia
information systems will result not only in a new approach to computer-based
anesthesia records but in a set of solutions that can be adapted
productively by other disciplines in the perioperative setting and
beyond."
The DDTF expects to have a preliminary reference
data set and a software tool for organizing this data available
for review in October 2002 at the Annual Meeting of the American
Society of Anesthesiologists.
The DDTF is supported through a cooperative effort
between industry and clinicians. Corporate participants in the DDTF
include: Draeger Medical, Deio, eko systems, inc., GE Medical Systems,
Philips Medical Systems, Picis, Inc., and Siemens.
The APSF is supported in part by a generous grant
from the American Society of Anesthesiologists.
Communication of important patient information
within and between disciplines has long been a topic of discussion
in medicine. With the advent of new technologies that facilitated
data collection and transfer -and expanded exponentially the quantity
of such data- concerns about effective communication increased.
These concerns became a focus of public attention in 1999, when
the National Academies of Sciences' Institute of Medicine (IOM,
Washington, DC) released a landmark study on medical error. The
authors of "To Err is Human: Building a Safer Health System" (1)
estimated that 44,000-98,000 Americans die each year as a direct
result of medical error--more than the total from motor vehicle
accidents (43,458) or breast cancer (42,297). Only the smallest
fraction of these accidents results from malfeasance or deliberate
sabotage of individual patients' well-being. Most are the result
of what the report called "systems problems," ie, failures in the
communication of accurate and/or timely information. Health care,
noted the IOM authors, lags almost a decade behind other high-risk
industries such as aviation in instituting basic measures that ensure
that the transfer to newer information technologies is accompanied
by safety and by direct and measurable benefit.
The IOM report challenged leaders throughout medicine
to "break the cycle of inaction" in dealing with medical errors.
With the increasing complexity of information management and transfer
systems, they said, these errors would only be compounded by inaction.
A national focus was needed "to create leadership, research, tools,
and protocols to enhance the knowledge base about safety."
The IOM report recognized that several groups
were already at work addressing these issues and named the APSF
as a leader in this field. The APSF is a nonprofit research and
educational organization dedicated to assuring patient safety. Formed
in 1985 with the mission that "no patient shall be harmed from anesthesia,"
APSF has been the leader in proactive, successful efforts to dramatically
improve the safety of anesthesia administration and reduce adverse
events. A key element in this success has been attention to early
identification of safety problems, promotion of research, dissemination
of information, and an emphasis on patient safety in clinical practice.
The APSF program has helped create a cadre of experts and a culture
and infrastructure devoted to promoting safety.
The APSF encourages the use of automated anesthesia
record-keeping systems to improve patient safety through a legible,
accessible patient record and the subsequent use of data collected
at the point of care for comparative research and performance comparison.
As a direct result of the IOM report and in support of its own ongoing
mission, the APSF held a workshop in October 2000 to discuss procedural
and patient safety issues involved in outcomes research and data
collection. The participants noted that, although anesthesia safety
has improved markedly over the past two decades, anesthetic mistakes
still occur and that a small percentage of these result in serious
morbidity and death. Assessment of the problem is challenging, because
anesthetic outcome data are primarily derived from retrospective
sources. A large-scale comparative study of outcomes and data to
support both broad and specific conclusions on causes of medical
errors in anesthesia is impossible without a unified "language."
The overall consensus of participants at the meeting was that the
APSF should be the leader in the development of a new and innovative
reporting system for anesthetic errors and outcomes.
The DDTF was formed in 2001 as a result of this
consensus (2). The long-term goal of the group is to develop an
active data dictionary that will direct database systems to obtain
outcome data and to link to computer-based anesthesia record keepers
throughout the country. The work of this task force and the commitment
of the APSF to the development of these tools should enhance knowledge
of perioperative outcomes, pinpoint sources of medical error and
suggest remedies, enrich research resources, and increase patient
safety.
Anesthesia information management systems (AIMS)
are currently installed in less than 3% of US hospitals, yet the
value of these systems is apparent in terms of legibility and accessibility
of the patient's medical record as well as for research and practice
analysis. According to another widely reported and influential release
from the IOM, Crossing the Quality Chasm (3), a move away from paper-based
documentation should be a top priority in establishing reliable,
comparable, and safe results in medical practice. The report also
noted:
- Automated clinical and administrative data also enable many
types of health service research applications, such as assessment
of clinical outcomes associated with alternative treatment options
and care processes; identification of best practices; and evaluation
of the effects of different methods of financing, organizing,
and delivering services.
- One of the most significant barriers to adoption of information
systems is the complexity and length of the product installation
process. Complete product installations lasting longer than 1
year are not uncommon. The result is a delay or barrier to (and
often initial bias against) the adoption of technology that will
improve the problems of legibility and accessibility associated
with paper-based anesthesia records. The installation of an AIMS
typically features the development of a customized set of terms
and phrases to be created by each institution. Few rules and minimal
standardization are available to guide institutions in developing
lists of terms.
This lack of standardization inhibits the sharing of data between
information systems from the same vendor and across institutions with
different vendor systems. Attempts have been made to standardize subsets
of terms pertinent to the specialty, such as the Association of Anesthesia
Clinical Directors' (AACD) Glossary of Times Used for Scheduling and
Monitoring of Diagnostic and Therapeutic Procedures (4), but these
lexicons have not been widely adopted and no one has attempted a more
complete terminology for the specialty.
Anesthesia has long recognized the need for aggregate
databases. A recent vendor-led attempt to create a national database,
the National Center for Clinical Outcomes Research (NCCOR), lacked
unified community support. NCCOR sought to create a virtual warehouse
to which participants would send data and from which NCCOR would
deliver comparative derivatives of their data in return. One of
the major obstacles to the effort was a lack of standard terminology.
Even those participating hospitals with AIMS faced the daunting
task of reformatting and normalizing their data into a standard
file for submission. Participating hospitals could not be sure that
their data collection was comparable with that of other hospitals
using data collections with slightly different terms and semantic
meanings.
The DDTF is approaching the communication problem
from a different perspective, recognizing that the primary problem
is the lack of a standard terminology. The technical arm of the
DDTF has been given the task of creating a tool for organizing a
lexicon of terms, based on a reference set. The clinical advisory
arm of the DDTF, the Clinical Working Group, under the direction
of Iain Sanderson, MD, Director of Anesthesia Informatics at Duke
University (Durham, NC), will use the tool to develop the actual
content of the reference set, pulling from existing standard terminology
wherever possible.
An immediate deliverable of this process will
be a set of standardized terms that could be preloaded into an information
system during an initial installation, providing a considerable
service to the vendors of these systems and the institutions that
invest in them. This baseline set of terms would not preclude customization
but ultimately could reduce the need for it.
The APSF intends to make the resulting reference
set of data available to the anesthesia community to encourage its
use in AIMS. AIMS manufacturers will benefit from a reduced cost
for research and market development. The anesthesia community benefits
from improved access to more cost-effective AIMS products that can
be more easily installed and with which data can be extracted for
evaluation and analysis of clinical practice and resulting patient
outcomes. The benefit would not be limited to those with automated
systems, however, as the experience of the AACD glossary has shown.
The terms would be useful in any research effort involving the aggregation
of anesthesia data from different institutional and research settings.
In response to the APSF's initiative to create
a data dictionary, Sanderson and colleagues have developed a 54-page
top-level design specification for a suite of tools called DATAMS
(Distributed Anesthesia Terms and Mapping System) that will create,
maintain, and disseminate a reference set of anesthesia terms and
phrases (5). The main design principle is the creation of a central
database of terms and term groupings that constitute the reference
set and a suite of applications that manage the data. Multiple users
distributed throughout the world may one day become involved--through
their desktops--in the creation and maintenance of the reference
set. DATAMS is designed to have a wide geographical usefulness bounded
only by the reach of the Internet.
DATAMS eventually will consist of a suite of applications
acting on a database. These can be divided broadly into performing
three discrete functions:
- Authoring and browsing a reference set of anesthesia terms;
- Providing a mapping tool that allows a third-party hospital
or vendor to map their terms against the reference set; and
- Providing tools to maintain a database of the reference set
and its revisions and of all mappings against it.
The DDTF Clinical Working Group envisions this
reference set of terms not as a static and unidimensional list but
as a dynamic database created through broad participation within
the global anesthesia community and "moderated" from a central source.
This model of involvement has been used successfully in other open-source
information projects, most notably in the Linux operating system.
One of the most exciting and potentially productive
results of central maintenance of the reference set is the potential
for individual institutions and vendors to "map" their custom terminologies
against the reference set, with DATAMS managing a brokerage of mapping
data. This approach will allow institutions and vendors to share
patient data by being supplied through DATAMS with the linking metadata
of their terminologies. Despite the high level of communication
this will ensure, privacy concerns will be avoided, because DATAMS
will not store patient data.
The design specification document discusses the
architecture of the new tool in detail. The core software platform
for DATAMS will be JAVA, which provides considerable platform independence
and the capability for multilingual support. The overall system
architecture makes use of a JAVA client application, local eXtensible
Markup Language (XML) data structures, a middle-tier layer of application
logic, and a database server. An extension to XML, called Reference
Set Markup Language (RSML), is proposed to manage the data transactions
involved in the DATAMS applications. Sanderson notes that although
commercial products similar to DATAMS are currently marketed, "a
proprietary tool lacks the full functionality envisioned by the
DDTF and would ultimately prove both limiting to the specialty of
anesthesia and quite expensive."
"We welcome this community-wide effort that brings
together the clinical expertise of the APSF, its expert clinician
panel, and the visionary manufacturers of anesthesia information
systems," says Stoelting. All those involved in this effort realize
that it will be challenging, time consuming, and require the ongoing
commitment of the broadest representation from all sectors of the
anesthesia community. The ultimate goal, however, is one that should
not only benefit anesthesia practice and research but serve as a
useful pattern on which other disciplines can base data dictionary
and standardization efforts. Most important are the direct results
in patient care. "The commitment of the APSF to the development
of these outcome tools should greatly facilitate knowledge of perioperative
outcomes and enhance patient safety," says Monk.
1. Committee on Quality of Health Care
in America, Institute of Medicine.. To Err is Human: Building a
Safer Health System. Washington, DC: National Academy Press, 2001.
2. Monk TG.. Any reporting system needs
postoperative outcomes and data collection. American Society of
Anesthesiologists Newsletter. 2001:65.
3. Committee on Quality of Health Care
in America, Institute of Medicine.. Crossing the Quality Chasm.
Washington, DC: National Academy Press, 2001: 182.
4. Donham RT.. Defining measurable OR-PR
scheduling, efficiency, and utilization data elements: the Association
of Anesthesia Clinical Directors procedural times glossary. Int
Anesthesiol Clin. 1998;36:15-29. 1998;36:15-29.
5. Sanderson, Iain.. Distributed Anesthesia
Terms and Mapping System (DATAMS) Software Requirement Specification.
Unpublished report: prepared for the Data Dictionary Task Force
of the Anesthesia Patient Safety Foundation 2002
For more information about the material contained
in this document, contact:
G. Rebecca Haines
Project Director
APSF Data Dictionary Task Force
PO Box 24
Paeonian Springs, VA 20129
703-475-7897
|
