Building the Right Patient-Centered
EMR
By W. Jördan Fitzhugh
The patient is the single most important source of information about his or her
problem. Designing an EMR that takes advantage of that knowledge requires some tough
choices.
|
|
W. Jördan Fitzhugh
|
All electronic medical records systems share the same basic requirements: reliability,
security and utility. But when patients are direct participants in the data collection
process, additional considerations become important, such as ease of use, increased
resiliency to errors and use of alternate languages for non-English speaking patients.
Adding the Internet as the method of delivery increases the scope and complexity
of these issues.
The design process also may also include decisions about whether to use browsers
for client access, how best to construct patient user interfaces, and how to make
sure information is collected quickly and securely. Here’s one approach to resolving
these issues.
Ease of Use, Update Trade-offs
Benefits of browser-based EMRs include almost universal familiarity with the browser
application, reducing the initial user learning curve; a reasonably standardized
set of client application capabilities, potentially reducing the programming task
for the implementation team; and an integrated encrypted communications capability,
which can help satisfy some security and privacy requirements.
On the other hand, implementation and support teams may not be able to control the
end user’s choice of browser version. Security deficiencies in browsers are discovered
with uncomfortable frequency. Finally, browser-based access mandates that most computation
be done at the server, increasing workload at the server and communications bandwidth.
Advantages in a dedicated client program are the freedom to offload most computations
to the client program, reducing server workload; complete control over all aspects
of the user interface; and enhanced security by employing multiple client-server
authentication protocols for each transmission. By providing automatic updates,
the support team’s need to stay current with multiple versions of software provided
by different vendors is eliminated. Because the EMR client application is used every
day by the end users, the small increase in the learning curve can be rapidly repaid
in gains in flexibility and reliability.
Shifting the Workload
In an outpatient setting, the caregiver’s three primary sources of diagnostic information
are the interview with the patient, the physical examination and the results of
laboratory tests. Of these, the most important is the interview. In fact, 70 percent
to 90 percent or more of the information to make a diagnosis comes from the patient’s
problem-specific medical history. However, the patient interview is often the most
difficult and most time-consuming portion of the office visit.
By involving patients as direct participants in their own health care, several advantages
accrue: with properly constructed problem-specific interviews, most of the workload
of capturing first-level information about the patient’s disease state is shifted
from caregivers to patients. Because the standardized patient interview is conducted
at the patient’s own pace, on the patient’s own time, prior to the face-to-face
encounter with the physician, the physician is able to recover some of the time
previously used in the interview process.
The content of the problem-specific interview is a medical issue that involves IT
teams in an ongoing process of review, maintenance and update. Tracking a specific
patient encounter to the exact version of the medical interview in use at the time
is critical for data reconstruction, auditing, accountability and validation.
The patient’s user interface must be as simple as possible, require no presumption
of computer skills, and be intuitive and non-intimidating. For example, simple text
displays can be used on the screen with numbered question responses. The patient
could select responses by pressing number keys from one to six, corresponding to
statements that apply, so no computer literacy or mouse facility is required. Patients
are immediately comfortable with pressing numbers on the computer keyboard and require
only a few seconds of instruction by a clinical staff person before being allowed
to complete questionnaires at their own pace.
The interface also should support multiple languages. This involves a combination
of two techniques. First, every discrete piece of medical information used in the
system is identified and assigned a coded phrase number. Each phrase is translated
and reviewed in context to ensure translation accuracy. During system operation,
a single keystroke switches between primary and alternate languages.
Encounter reports can be constructed in bullet fashion, using response phrases associated
with questions seen by the patient, with no attempt to artificially synthesize paragraphs.
Consequently, the language can be switched at any time to the reader’s preferred
language, even after the patient encounter has been completed, finalized and locked
from changes.
Enabling Privacy, Security, Auditing and Research
Design challenges concerning information security, privacy, change auditing and
authorized medical data transmission are amplified when using the Internet as a
global communications medium, with users permitted access from anywhere in the world.
One specific issue is that of allowing users to view information in local time while
preserving the chronological sequence of events for audit and data reconstruction.
One design solution uses Greenwich Mean Time for the data center’s common time,
with adjustment back to local time by the client program where necessary. To enable
data extracts for auditing or to create research databases while maintaining patient
privacy, content-free identifiers (globally unique identifiers, or GUIDs) are used
to link related information within the data center. Each patient, while having a
chart number and/or personal identifier assigned for use by the end user, is also
assigned a GUID, which internally links the patient’s personally identifying information
with encounter information such as diagnoses, treatments and consultations.
This data structure allows the creation of secondary databases containing properly
linked medical data but prevents tracking back to the individual patient. Using
GUIDs as internal identifiers also permits combining patient charts when the same
patient is found to have been entered into the system more than once, and allows
re-splitting the combined chart if such a combination is later found to have been
made in error.
All of these issues are essential to consider when patients are the direct source
of EMR data collection. By reducing the complexity of the system and increasing
performance, capability and security, far more control is obtained over all aspects
of the EMR.
W. Jördan Fitzhugh, M.S., is executive vice president
of PrimeCare Systems Inc., Newport News, Va.
Reprinted with the permission of HealthForum
and HHN Most Wired Magazine.
View the original article
.
For questions or comments regarding our web site, please send e-mail to
webmaster@pcare.com.
This page, and all contents, are Copyright (C) 1998-2008 by PrimeCare Systems, Inc.