Cases & Commentaries

Electronic Err

Commentary By Paul C. Tang, MD

The Case

A 75-year-old woman with coronary artery disease
presented to the emergency department (ED) with chest pain that
that had not responded to three sublingual nitroglycerin tablets at
home. Supplemental oxygen and nitroglycerin paste resulted in
resolution of the patient's symptoms, but she was admitted for
cardiac monitoring and serial cardiac enzymes to rule out
myocardial infarction.

The patient gave a clear history to the admitting
internist, but could not recall the names of some of her
medications, nor could she remember any of the doses. The hospital
was able to access the electronic health record (EHR) of the large
multi-specialty clinic where the patient received her medical care.
The admitting physician printed the medication list from that EHR.
The most recent note in her ambulatory chart listed warfarin,
aspirin, clopidogrel, diltiazem CR, metoprolol XL, and
atorvastatin—so he wrote orders for those medications with
doses as stated in the medication list.

One hour later, the admitting physician received
a page from the telemetry floor. The nurse informed him that the
patient had developed a junctional rhythm with a heart rate less
than 40 and stated, "She looks really bad." She was given atropine
stat, which resulted in improvement in the heart rate and the
patient's general appearance.

On reviewing the patient's outpatient record in
greater detail, the physician found a recent cardiology note. The
note listed the same medications documented elsewhere in the EHR,
but also documented his plan to discontinue diltiazem and decrease
metoprolol due to recent episodes of symptomatic bradycardia.

Neither the EHR medication list nor the most
recent note from the primary care physician reflected these
changes, but a call to the patient's pharmacy confirmed that she
had not refilled her diltiazem and that metoprolol had been
prescribed at a lower dose than before. Thus, the patient had been
given 100 mg of metoprolol XL and 180 mg of diltiazem CR by the
hospital admitting physician, rather than the 50 mg of metoprolol
XL and no diltiazem intended by her outpatient cardiologist. The
patient remained clinically stable, but did rule in for a
myocardial infarction with a troponin that peaked at 8. Whether the
infarct resulted from the medication error or had already occurred
at the time of admission was unclear, but the physician did inform
the patient of the error.

The Institution's Response

After a discussion of this case at the clinic's
monthly safety and quality improvement meeting, a physician and
nurse audited a small random sample of patient records. Medication
lists commonly lagged far behind clinic notes, frequently
containing medications no longer received by the patient, omitting
new medications, and failing to document changes in dose. Analogous
discrepancies were found for problem lists: patients' main EHR
problem lists often continued to list problems long since resolved
and failed to include new problems.

The clinic approached the EHR vendor to ask about
modifications to the current system that would allow automatic
updating of medication and problem lists or, in the absence of such
a system, add a flag to medication and problem lists indicating
that the patient has been seen in clinic more recently than the
date of the last change to the medication list. The vendor thought
that the latter solution could be implemented over a 6- to 12-month
time frame. In the meantime, clinic nurses added a medication
review to vital signs and weight prior to each patient's
appointment with a physician.

The Commentary

This interesting case raises the question whether
use of an electronic health record (EHR) system can create new
opportunities for medical error that were not present in the paper
system. Errors can arise either because the new technology exposes
underlying latent
errors that had not been visible previously or the new
technology itself introduces opportunities to make mistakes. Errors
that result from using the new technology can be caused by the
workflow required to use the technology (eg, a cumbersome user
interface may cause a user to pick the wrong item; poor methods of
presenting data can obscure important information) (1) or from misinterpretation or misuse of the output from
the technology.

In this case, the error is one in which the new
technology exposes deficiencies in existing record-keeping
procedures—the existence of the EHR made the inaccurate
information contained in the medical record more readily available.
Regardless of whether EHR use exposes deficiencies in the
underlying paper system or creates new error opportunities,
however, both situations are dangerous to patient care and need to
be better understood as the country proceeds in wide-scale
implementation of EHR systems.(2)

The medical error, and subsequent adverse event,
reported in this case arose when the admitting physician based
medication decisions on the content of the medication list
contained in the EHR system used by the patient's primary care
physician. Unfortunately, a specialist changed the patient's
medications and documented the change in the progress note, but did
not update the active medication list. Hence, the patient's medical
record contained inconsistent information. In the EHR system
described, maintenance of the medication list was a manual task, as
it is in the paper system. Consequently, the inconsistency could
have occurred in either a paper-based medical record or an EHR.

Many of the current comprehensive EHR systems
automatically populate the active medication list when electronic
prescriptions are written or renewed. When changes to doses are
made or a medication is discontinued, however, the physician must
actively update the existing medication list.

The accuracy of information contained in paper
records has been studied more extensively than the accuracy of data
in EHR systems. Results from these studies of paper records have
been disturbing. In one study, 40% of encounters had no diagnosis
recorded and 30% had no therapeutic agent recorded.(3) In another study, up to 49% of the visits did not
include a well-defined treatment plan in the record.(4) Perhaps one reason the status quo has been maintained
so long is that the deficiencies have been hidden by the
unavailability or illegibility of the paper record.

Although there are fewer studies of the accuracy
of information in EHR systems, the results are more encouraging,
although still far from ideal. In one study, medication lists were
found to be 83% correct and 93% complete.(5,6) In a study assessing the completeness of medication
lists maintained by physicians using an EHR compared with
medication lists maintained by physicians using a paper record in
the same department, the medication lists in the EHR were judged by
a blinded expert panel to be approximately twice as complete as
those in the paper record (p7) Regardless of which media is used to store the
medication list, its dependence on manual updates when medications
are changed leaves room for error. And, as was true in this case,
the situation is further complicated when multiple parties are
managing a patient's medications.

Methods to encourage physicians to
conscientiously maintain up-to-date medication lists have
traditionally relied on education and peer pressure. When using an
EHR system, several factors can encourage or reward physicians for
maintaining summary information such as the active medication list.
In an electronic record, one of the main incentives to maintain an
accurate summary list is that it allows for instant access to
information, something that cannot be replicated in the paper
world. Of course, the technology can help facilitate keeping
accurate and up-to-date summary lists. For example, in a properly
designed EHR, writing an electronic prescription will automatically
update the summary medication list. Moreover, the act of writing an
electronic prescription can automatically make an entry in the
progress note, in contrast to the double entry necessary with paper
records.

Although there is a great deal of hope and
optimism that more widespread use of EHR systems will significantly
improve patient safety, this case should remind all clinicians that
having instant access to information does not guarantee its
veracity. The traditional best practice of using multiple sources
to verify the accuracy of information critical to making key
clinical decisions (eg, cross checking summary lists with other
sources of information) is still warranted. The advantage of using
an EHR system is that more complete information is available
anytime. In a blinded expert-panel study assessing medical
decisions of physicians using an EHR compared to those of similar
physicians using traditional paper records, the experts scored the
EHR-supported decisions as significantly more appropriate than
those of physicians using a paper record.(7)

Another patient-safety policy that could be
incorporated into routine practice with EHR systems would be to
produce an up-to-date printout of the current medication list for a
patient at the end of each encounter. Creating such a list is a
JCAHO requirement for hospital discharge. It makes just as much
sense in the outpatient setting, and in our experience at the Palo
Alto Medical Foundation, patients greatly appreciate it. This
practice has the additional benefit of further motivating
physicians to keep patients' medication lists updated.

When implementing EHR tools to improve access to
and management of clinical information, it is important not to
abandon the safety practices used in the past, but rather to use
the richer set of patient safety tools provided by the EHR system
to continuously improve the safety performance of the delivery
system. The safety practices described in this case study (eg,
informing the patient of the error, assessing the root cause of the
adverse event, performing a hazard analysis, and developing
prevention strategies) indicate that a healthy culture of safety is
in place to respond to safety incidents. Use of EHR safety tools
will go a long way toward preventing them from occurring in the
first place, as espoused by the Institute of Medicine
(IOM).(8)

As the country embarks on initiatives to increase
the adoption of electronic prescribing and EHR systems,
practitioners should keep in mind that these computer systems are
merely tools that must be used effectively by physicians to reduce
the number of medical errors.

Providers should keep the following priorities in
mind when acquiring or using EHR systems:

  • Select electronic health record
    systems that incorporate the IOM's key capabilities (9) addressing patient safety and medication safety (eg,
    drug interaction checking, rule-based alerts, preventive services
    reminders, and patient access to the EHR).
  • Increase physicians' awareness of the benefits of
    maintaining an accurate medication list, and the liabilities of not
    doing so, through education and by using case studies such as this
    one.
  • Institute practices
    to provide patients with an up-to-date list of current medications
    and instructions after every encounter.
  • Continue to use best practice safety techniques
    (eg, verifying critical patient information) when making key
    clinical decisions based on information in the medical
    record.

Paul C.
Tang, MD
Chief Medical Information Officer
Palo Alto Medical Foundation

References

1. Ash JS, Berg M, Coiera E. Some unintended
consequences of information technology in health care: the nature
of patient care information system-related errors. J Am Med Inform
Assoc. 2004;11:104-12.[ go to PubMed ]

2. Thompson TG, Brailer DJ. The decade of health
information technology: delivering consumer-centric and
information-rich health care. July 21, 2004. U.S. Department of
Health and Human Services Web site. Available at: [ go to related site ]. Accessed September 16,
2004.

3. Dawes KS. Survey of general practice records.
BMJ. 1972;3:219-23.[ go to PubMed ]

4. Tufo HM, Speidel JJ. Problems with medical
records. Med Care. 1971;9:509-17.[ go to PubMed ]

5. Wagner MM, Hogan WR. The accuracy of medication
data in an outpatient electronic medical record. J Am Med Inform
Assoc. 1996;3:234-44.[ go to PubMed ]

6. Hogan WR, Wagner MM. Accuracy of data in
computer-based patient records. J Am Med Inform Assoc.
1997;4:342-55.[ go to PubMed ]

7. Tang PC, LaRosa MP, Gorden SM. Use of
computer-based records, completeness of documentation, and
appropriateness of documented clinical decisions. J Am Med Inform
Assoc. 1999;6:245-51.[ go to PubMed ]

8. Aspden P, Corrigan JM, Wolcott J, Erickson SM,
eds. Patient safety: achieving a new standard for care. Washington,
D.C.: The National Academies Press; 2004.[ go to
related site
]. Accessed September 16,
2004.

9. Institute of Medicine. Committee on Data
Standards for Patient Safety. Key capabilities of an electronic
health record system. July 31, 2003. Available at: [ go to
related site
]. Accessed September 16,
2004.