View as “PDF version”
Patient Safety Tip of the Week
March 12, 2024
Double Checks
Don’t Do It Again
There’s a reason that double checks don’t make the highly
effective level in the hierarchy of
effectiveness of interventions – they don’t always work!
In our October 16,
2012 Patient Safety Tip of the Week “What is the Evidence on Double Checks?” we reviewed some of the evidence on double
checks. Our conclusion was that double checks remain a relatively weak safety
intervention and they are prone to errors but, done correctly, the independent
double check probably does provide an additional element to our defenses
against errors. We recommended that you audit those processes for which you
require double checks, see how often you are actually doing
them and doing them correctly and truly independently, and then put some
structure into your double check process (eg.
checklists, forcing functions, etc.). That column and the ones listed below
have urged caution in the use of double checks.
Now a new study (Konwinski 2024) analyzed data on independent double checks done on high
alert medications in a 24-bed pediatric ICU. They examined 37,968 high-risk
medications administrations to 4417 PICU patients over a 40-month period.
Compared to those instances using single checks, those using independent double
checks did not lower the rates of medication administration events, hospital
length of stay, or patient mortality. Each double check took an average of 9.7
minutes, and a single check took an average of 1.94 minutes. Nursing
favorability for single checking increased from 59% of nurses in favor during
the double check phase, to 94% by the end of the single-check phase. The
researchers conclude that performing independent double checks on high-risk
medications administered in a pediatric ICU setting afforded no impact on
reported medication events compared with single checking and was quite
inefficient.
Konwinski and colleagues had actually
reported on this study earlier (Konwinski
2021) and nicely described how they got there. Their hospital had done 4
improvement projects looking at adherence to the independent double check
process without seeing any reduction in medication events. So, they analyzed
their process using SEIPS (Systems Engineering Initiative for Patient Safety) methodology.
Once they identified the barriers and inefficiencies of their independent
double check system, they decided to do a pilot study of single-checking these
high-alert medications.
Pediatrics is one area in which you’d
think double checks would be wise. In pediatrics, medication doses are often
calculated based upon patient weight or, in some cases, body surface area. Any
time you have to do such calculations, the possibility
of error is introduced. Hence, the concept of doing truly independent double
checks (where 2 individuals independently do the calculations and then compare
their results) is a logical one. But we suspect several factors probably played
a role in the Konwinski study’s failure to show a positive impact of
independent double checks. Most importantly, computers are now likely doing all the calculations and they are probably much better
at doing these calculations than humans are. Secondly, this study was in an era
where barcode scanning is actually often performing
the equivalent of a double check. So, these technology
factors have reduced the need to perform double checks in many circumstances.
But we can’t get complacent. The “mental
double check” is still important. The attitude that “the computer can’t be
wrong” can lead to failure to prevent some serious accidents. Barcoding systems
and computer calculations are only as good as the data entered. We always give
the example where a physician erroneously ordered 100 units of regular insulin
instead of the 10 units the patient was actually supposed
to receive. The nurse scanned the medication and the patient’s ID bracelet and the computer told her that was the right
patient, right medication, right dose, and right mode of administration. The
nurse did a delayed “mental double check” after she administered the insulin
and became concerned that this was an incorrect dose for that patient. She drew
blood for a stat glucose and gave a bolus of D50W. The
blood glucose confirmed hypoglycemia and review of the patient’s outpatient
records confirmed his usual dose was just 10 units. Another error avoided by
the “mental double check” is when a patient weight is entered into the computer
in pounds rather than kilograms and the computer calculates a much higher dose
than intended. Konwinski et al. do note that their nurses can still ask another
nurse to double check when they have questions.
So, should we abandon the
independent double check? There is certainly a reduced need for double checks
in most circumstances. But there are others where they should be used. We go
back to the Vanderbilt incident where there was inadvertent administration of a
neuromuscular blocking agent (NMBA) with a fatal outcome. In that case, there
was no barcoding capability in the area in which the nurse was to administer a
medication. Hence, the “double check” that would have been done via barcoding
was unavailable. Another example where we think independent double checks can
be valuable is in programming infusion pumps that are infusing some
chemotherapy agents (though we’ve even described cases where such infusion
accidents occurred despite supposed double checks!).
Note that in last week’s column (our March 5, 2024 Patient
Safety Tip of the Week “2
ISMP's Update High Alert Medication Lists”) we noted both ISMP and ISMP
Canada have updated their high-alert medication lists. Both ISMP’s still
mention independent double checks as potential strategies to avoid errors with
high-alert medications. But both stress that use of double checks should be
limited. ISMP Canada states “independent double checks can be useful, but a requirement for too many
independent double checks may introduce ‘checking fatigue’. Rather, such
double checks should be implemented at the most impactful point(s) in the
medication-use process and should be supplemented by other effective
strategies.” Good advice.
Konwinski and colleagues are to be
commended for challenging the status quo and questioning “that’s the way we’ve
always done it”. We hope that you’ll also go back and read some of our prior
columns that have addressed the issue of how well (or not so well) double
checks work.
Some of our other
columns on double checks:
January 2010 “ISMP
Article on Double Checks”
October 26, 2010 “Confirming
Medications During Anesthesia”
October 16, 2012 “What is the Evidence on Double Checks?”
December 9, 2014 “More
Trouble with NMBA’s”
April 19, 2016 “Independent
Double Checks and Oral Chemotherapy”
December 11, 2018 “Another
NMBA Accident”
January 1, 2019 “More
on Automated Dispensing Cabinet (ADC) Safety”
March 5, 2019 “Infusion
Pump Problems”
August 27, 2019 “Double
Check on Double Checks”
November 19, 2019 “An
Astonishing Gap in Medication Safety”
April 14, 2020 “Patient
Safety Tidbits for the COVID-19 Pandemic”
March 2020 “ISMP
Smart Infusion Pump Guidelines”
August 4, 2020 “Intravenous
Issues”
August 18, 2020 “More
Caution on Double Checks”
References:
ISMP Canada. Hierarchy of Effectiveness. ISMP Canada 2013
https://ismpcanada.ca/resource/hierarchy-of-effectiveness/
Konwinski L, Steenland C, Miller K,
et al. Evaluating Independent Double Checks in the Pediatric Intensive Care
Unit: A Human Factors Engineering Approach. Journal of Patient Safety 2024; ():10.1097/PTS.0000000000001205,
January 18, 2024
Konwinski L, Miller K, Steenland C.Leveraging Single-Checks to
Improve Medication Safety. An examination of how humans interact with their
environments and each other led this team to question one of its long-standing
medication safety practices and change how they work. Children’s Hospital
Association 2021; Published July 23, 2021
Print “PDF version”
http://www.patientsafetysolutions.com/