The road less traveled
From the Augus 21, 2002 issue
Isn't it frustrating when you think you've solved a problem,
but it reappears months later? There's a reason for this unsettling
sense of déjà vu. Simply put, the action taken
was ineffective, even if it seemed to solve the problem temporarily.
Then it's back to the drawing board to find a more effective
solution. But on this long and sometimes discouraging journey
toward solving problems, our initial steps may show results
that are partly effective; steps that clearly reduce the risk
of error, even if all chance of it happening has not been eliminated.
An error report submitted by a hospital illustrates this point
Midazolam syrup was used frequently on the hospital's pediatric
unit, so bulk bottles of the syrup (2 mg/mL) were stocked
in the automated dispensing cabinet. After several calculation
errors led to overdoses, the screens on automated dispensing
cabinets were redesigned to provide dose warnings. Dose conversion
charts also were posted nearby to eliminate the need for calculations.
This seemed to work for quite a while, but recently another
overdose occurred. A child was to receive 6 mg of midazolam.
The nurse had used the conversion chart to determine the correct
volume of syrup to administer (3 mL). However, a mental slip
caused her to confuse the dose in mg (6) as the volume to
be administered. She incorrectly drew up 6 mL of syrup and
administered 12 mg of midazolam to the child. Luckily, no
Certainly more could be done to prevent overdoses with midazolam
syrup. Unit dose dispensing of midazolam doses from the pharmacy
is just one example. But more to the point, this error report
clearly typifies where many healthcare providers are today
- not at the final destination of error prevention,
but on the right road, with frequent pit stops to try out
new error reduction strategies. And it's good to be
on the right road, heading in the right direction. But as
we improve processes and systems, errors are still possible
if people are involved. Mental slips and lapses can happen
to the most experienced, vigilant, and well-trained clinicians
- they are unintended and unpredictable, thus largely unmanageable.
As a result, human error should be anticipated and methods
to detect errors must be employed to prevent serious harm.
While technological solutions such as computerized prescriber
order entry and bar coding systems have great potential to
detect human error, manual redundancies such as independent
double checks still play an important role in error detection.
Studies show that manual redundancies detect about 95% of
errors. And while the 5% of errors that get through human
detection systems make it unreliable as an error prevention
strategy, that's hardly the point if a manual check is the
only way to detect an error that could lead to patient harm.
In the end, independent double checks serve two purposes:
to hopefully, though not dependably, detect a serious error
before it reaches a patient; and just as important, to bring
attention to the systems that allow the introduction of human
error. Using the example above, if an independent double check
had been employed before giving the child midazolam syrup,
there was a 95% chance that the error would have been detected.
If it did, an equally important step would be to bring the
issue back to the table for discussion on how to further reduce
the risk of an error or prevent it. This takes error
prevention to the next level, rather than relying on
an imperfect but necessary manual check system. Likewise,
if an independent double check had been employed in the case
above, but the error hadn't been noticed until after the child
received the overdose, the issue wouldn't be the failed double
check. More importantly, the issue still would be how to fix
the system so that we can move further along the road toward
error prevention. In the end, by taking this more difficult
path - the road less
traveled - we will be able to say, "and that has made
all the difference."