Misidentification of alphanumeric
characters
From the January 12, 2000 issue
PROBLEM: It's not uncommon to read a letter or number
differently than the writer intended. Recently, while reviewing
a handwritten, faxed order, a pharmacist read the word "IODINE"
in the space for allergy alerts. Yet, a second pharmacist
read the allergy as "LODINE." The prescriber was contacted
for clarification, and she identified LODINE (etodolac) as
the drug to which the patient was allergic. In another case,
a patient listed Lodine 400 mg BID as a medication she was
taking prior to admission. However, the admitting resident
misread the notation and wrote an order for saturated solution
of potassium iodide 400 mg BID. The patient received two doses
before the error was discovered.
Computerized physician order entry (CPOE) can overcome most
problems with poor handwriting. However, even typed or computerized
physician orders may not help prevent all of them. Anyone
familiar with e-mail knows how easy it is to misidentify a
computer-generated lower case letter L (l) in an e-mail address
as the numeral one (1), or the letter O as a zero (0)! Even
when using character recognition software, drug names may
be translated incorrectly. For example, when we tested Lodine,
typed with a lower case L, the software recognized the drug
name as Iodine. Likewise, it's easy to confuse the upper case
letter Z with the number 2. In fact, research conducted by
Bell Laboratories found that some symbols are more vulnerable
than others to misidentification1.
The previously mentioned characters (I/1; O/0 and Z/2) plus
the number 1, which can look like a 7, accounted for over
50% of the errors caused by character misidentification in
the study. Further, the context in which the order is being
read may not always be helpful in properly identifying alphanumeric
characters. While it would be unlikely to read ZETAR as "2TAR,"
it would be easy to read an order for "HCTZ50mg" as either
hydrocortisone 250 mg or hydrochlorothiazide 50 mg.
SAFE PRACTICE RECOMMENDATION: Although CPOE is on
the horizon (see article on page 2),
for now, handwritten orders are still the norm. Many drug
name recognition errors can be reduced with block printing
using upper case characters. Prescribers may save time by
using cursive writing, but this must be balanced against the
risk of error and the tremendous waste of precious staff time
when poorly handwritten orders must be interpreted. In the
absence of computer orders, outpatient prescription forms
should be formatted so that prescribers must print the name
and strength in designated blocks for each letter.2
The rules outlining each block should be preprinted in a 30%
shade so that they appear a light but visible gray. The lightly
outlined rules would help prevent confusion between a T, 7,
and I, or an E, F, and L. For inpatient order forms, rules
also can be shaded similarly, making them visible to prescribers,
yet light enough to prevent interference with characters when
reading or faxing. Back copies of order forms (NCR forms)
should never have lines.
Symbolic differentiation is another way to distinctively
convey a symbol's meaning.1 Throughout Europe, it's common
to see a zero written with a slash through it to differentiate
it from the letter "O." The number 7 can be written with a
bar through it to prevent confusion with the number 1. The
letter "Z" with a bar through it also can prevent confusion
with the number 2. Stricter adherence to these principles
would help reduce character misidentification.
The potential for name-related errors is greatly reduced
when pharmaceutical manufacturers incorporate practitioner
testing of names before trademarks are submitted to FDA. The
Institute of Medicine recently listed pre-market testing of
pharmaceutical trademarks as one of their official recommendations.
References: 1) Nierenberg GI. Do it right
the first time. New York: John Wiley and Sons 1996; 2) Davis
NM. Drug names that look and sound alike. Hosp Pharm 1999;34:1160-78.
|
