You are on page 1of 5

USP Medication Safety Forum

Medication Errors John P. Santell, M.S., R.Ph.


Involving Wrong Diane D. Cousins, R.Ph.

Administration Department Editor


James G. Stevenson, Pharm.D., College of Pharmacy,
University of Michigan, and University of Michigan Hospitals.

Technique This department features medication error issues based on


data collected by the United States Pharmacopeia (USP).

eports submitted to USPs MEDMARX database from A through I. Categories B through D include errors

R indicate that wrong administration technique is


the type of medication error that most often
results in harm to patients. Wrong administration tech-
that occur but do not cause harm, and categories E
through I include harmful or fatal (category I) errors.
Errors in categories G through I are sentinel events. (The
nique errors result from an incorrect and/or improper Joint Commission definition of a sentinel event is any
procedure used in the administration of a drug. Examples unexpected occurrence involving death or serious physi-
of such errors are shown in Table 1 (page 529). cal or psychological injury, or the risk thereof.)2
Of the 14 different types of error reported to MED- Of the 7,205 errors involving wrong administration
MARX, wrong administration technique errors are infre- technique, two resulted in death, 11 necessitated inter-
quently reported but result in a disproportionate number vention to sustain life, 8 may have resulted in permanent
of harmful or fatal outcomes. From 1999 through 2003, harm, 55 necessitated prolonged or intensified treat-
7,205 medication errors involved wrong administration ment, and 432 necessitated intervention of some type
technique, of which 508 (7%) were harmful. In compari- (total of 508 harmful errors).
son, only 1.84% of all errors reported to MEDMARX in Because of their harmful outcomes, wrong administra-
the same five-year period were harmful. tion technique errors affect the patients level of care,
Errors that involve wrong administration technique often resulting in the initiation of or increase in patient
were paired with other types of errors. In 1,257 cases, observation and vital signs monitoring, initiation of or
wrong administration technique was most frequently change in drug therapy, and initiation of laboratory test-
coupled with improper dose/quantity, followed by omis- ing. Also reported were administration of an antidote,
sion error, prescribing error, wrong route, wrong drug transfer to a higher level of care, administration of oxy-
preparation, wrong time, and unauthorized drug. gen, and extension of the hospital stay by one to five days.

Findings Location Where Errors Originated


Severity of Error in Terms of Patient Outcomes The reported location of more than half (56%) of the
The MEDMARX program enables subscribing hospitals administration technique errors was a nursing (patient
and health systems to track errors in a standardized format care) unit. Other frequently reported locations include the
and compare themselves with this nationwide database. emergency department (6%), medical intensive care unit
MEDMARX categorizes errors according to the National (ICU; 4%), and outpatient clinic (3%). Coronary, neonatal,
Coordinating Council for Medication Error Reporting and and surgical ICUs; pediatrics; and oncology each were
Prevention severity index.1 The index has nine categories mentioned as locations of the error in 2% of the reports.

528
September 2005 Volume 31 Number 9
Copyright 2005 Joint Commission on Accreditation of Healthcare Organizations
Table 1. Examples of Errors Involving Wrong Table 2. Common Causes of Medication
Administration Technique Administration Technique Errors Reported to
MEDMARX*
Not activating the drug product chamber into the
main intravenous (IV) solution chamber to produce Cause of Error %
a thorough mixture Performance deficit 45
Crushing sustained-released tablets Procedure/protocol not followed 30
Placing a patients eyedrops in the wrong eye Knowledge deficit 27
Administering an IV medication too rapidly Monitoring inadequate/lacking 9
Not flushing an IV line Communication 7
Pump, improper use 7
During the administration process, mixing a drug
in an incompatible solution so that a precipitate Dispensing device involved 6
forms
* Based on 12,145 causes of error identified in 6,885 MEDMARX records
for 1999 through 2003.

Reported Causes and Contributing Factors


For the administration technique errors reported in A nurse administered undiluted potassium phosphate
1999 through 2003, 12,145 causes of error were identified injection by rapid IV push. The patient went into cardiac
(Table 2, above). In many cases, more than one cause arrest. Resuscitation attempts were unsuccessful.
was reported. Performance deficit was the most com- A patient receiving IV vancomycin developed red man
monly reported cause of error and is essentially defined syndromelightheadedness and an erythematous rash
as a cause that results when the person involved had the on the face, neck, chest, and upper armsafter rapid
requisite knowledge and training to perform the neces- administration of a dose given in just several minutes, a
sary task but failed to do so. The most frequently identi- practice that was not consistent with the institutional
fied contributing factors to administration technique protocol for use of the drug.3 The rate of IV infusion of
errors were distractions, inexperienced staff, and work- 0.9% sodium chloride injection (normal saline) was
load increase. increased to correct the patients hypotension. The symp-
toms resolved within several hours and did not recur
Medications when subsequent vancomycin doses were infused for one
Of the 508 reports of harmful administration technique hour, the rate specified in the institutional protocol.
errors, 442 identified a drug product (Table 3, page 530). A patient in the neonatal ICU experienced asystole
after IV tubing containing residual potassium chloride
Actions Taken solution was flushed too rapidly (within 10 minutes).
Many of the reports of administration technique errors Calcium gluconate was administered, and the infant was
identified actions that were taken to avoid similar errors, successfully resuscitated.
such as informing the staff member who made the initial An order was written for phytonadione 1 mg IV. The
error (62%), providing education or training (31%), pharmacy dispensed a 5-mg vial of the drug instead of
informing staff who were also involved in the error (20%), diluting the drug in preservative-free 5% dextrose injec-
and enhancing communication (12%). Five percent of the tion or 0.9% sodium chloride injection for IV infusion at
reports said the patient or caregiver was informed, and a rate not to exceed 1 mg/min in accordance with estab-
4% said the patients physician was informed. lished procedures. A 1-mg dose was prepared on the
nursing unit using sterile water for injection containing
Case Reports parabens (preservatives). It was given by rapid IV push,
The following case reports, which are based on and the patient experienced dyspnea, wheezing, chest
records in the MEDMARX database, highlight common discomfort, and electrocardiographic changes. The
administration technique errors: patient was treated with several medications to counter

529
September 2005 Volume 31 Number 9
Copyright 2005 Joint Commission on Accreditation of Healthcare Organizations
Table 3. Leading Products Associated with pain was not well controlled. The dose of the morphine
Wrong Administration Technique Errors PCA was increased several times but the medication did
Resulting in Harm Reported to MEDMARX* not provide relief. After 10 days, nursing staff discovered
that the morphine PCA had been incorrectly loaded into
Generic Drug Product Name Ranking
the pump and that the infusion never entered the
Morphine 1
patients body. The error was corrected and the patient
Heparin 2 began receiving the increased (last ordered) dosage of
Dopamine 3 morphine. The patient shortly thereafter went into respi-
Vancomycin 4 ratory depression and was administered naloxone.
Phenytoin 5 A physician ordered phenytoin 1,500 mg IV for 2
Potassium chloride 6 hours. The pharmacy sent six vials containing 250 mg/5
Hydromorphone 7 mL to the nursing unit. The nurse diluted the phenytoin
Fentanyl 8 in 100 mL of 0.9% sodium chloride injection and infused
the admixture into a wrist vein. Severe necrosis of the
Insulin 9
hand developed, resulting in an extended hospital stay
* Based on 406 reports of administration technique error involving and eventual loss of the smallest finger. Phenytoin sodi-
patient harm and where a product was identified.
Designated as a high-alert medication based on the definition devel- um may be administered by direct IV injection into a
oped by the Joint Commission on Accreditation of Healthcare large vein through a large-gauge needle or catheter.5 IV
Organizationsmedications that have the highest risk of causing injury if
misused. MM.7.10: The hospital develops processes for managing high- infusion of phenytoin is not generally recommended
risk or high-alert medications [MM-19]. Source: Joint Commission on because of the danger of precipitation, but some clini-
Accreditation of Healthcare Organizations: 2005 Comprehensive
Accreditation Manual for Hospitals: The Official Handbook. Oakbrook cians suggest that this route is feasible if the infusion
Terrace, IL: Joint Commission Resources, 2004. fluid is suitable, as it was in this case; the drug is suffi-
Based on the number of times the drug product was reported.
ciently diluted (< 6.7 mg/mL, compared with 15 mg/mL in
this case); a 0.22- inline filter is used; and the admixture
the negative effects and recovered. It was unclear is administered immediately after preparation and
whether the symptoms were the result of the drug (that observed during infusion.
is, hypersensitivity, anaphylaxis), the diluent (only pre- In other cases, local phlebitis developed in a patient
servative-free 5% dextrose or 0.9% sodium chloride is receiving an IV phenytoin admixture through a Y-site
recommended by the manufacturer), or the rapid admin- connection into tubing containing 5% dextrose injection,
istration rate.4 with which phenytoin is incompatible. An inline filter
An infant receiving IV fentanyl by syringe pump devel- was not used, despite labeling on the phenytoin contain-
oped bradycardia and dusky skin. Positive-pressure ven- er with instructions for use of such a filter.
tilation had little effect. Rigidity was noted in the infant, Propranolol was given without checking the vital
and the nurse noticed that the fentanyl syringe contained signs in a patient who had been experiencing hypoten-
a smaller volume than it should have. The fentanyl infu- sion (systolic and diastolic blood pressure of 85 mmHg
sion was discontinued, naloxone was administered to and 60 mmHg, respectively) and lethargy. The patient
reverse the effects of the opiate, and oxygen saturation experienced orthostatic hypotension and fell when
improved immediately. Investigation revealed that the attempting to get out of bed.
syringe pump had been mistakenly programmed to deliv- A nurse crushed a 40-mg extended-release oxycodone
er the fentanyl solution at a rate of 1 mL/hr instead of 0.1 tablet (extended-release tablets should not be crushed
mL/hra rate that was 10 times what was ordered. This because crushing causes immediate release of the drug)
is an example of two types of errorwrong administra- and administered it to a patient. The patient experienced
tion technique and improper dose quantity. respiratory arrest and was successfully resuscitated
A patient receiving morphine from a patient- after intubation and the administration of naloxone to
controlled analgesia (PCA) pump complained that his reverse the opiate effects.

530
September 2005 Volume 31 Number 9
Copyright 2005 Joint Commission on Accreditation of Healthcare Organizations
A heparin IV infusion was ordered in accordance with incompatible with dextrose 5% or must be used by _)
a dosing nomogram that called for a rate of 33 mL/hr. need to be clearly visible.
However, the infusion was initiated at a rate of only 22 On patient care units, preprinted charts of standard-
mL/hr. The error was discovered only because the ized infusion rates should be readily accessible and reg-
patients activated partial thromboplastin time (aPTT) ularly used.
was not within the therapeutic range. A therapeutic Infusion devices and dispensing devices (for example,
aPTT was achieved after the heparin infusion rate was devices for compounding parenteral nutrient solutions)
increased to 33 mL/hr. This case concerning a pump set- have become more numerous and sophisticated. The
ting, as in the earlier case of the infant given fentanyl proper use of IV infusion pumps and dispensing devices
with an improperly programmed syringe pump, involves should be addressed in staff training programs. A review
not only wrong administration technique but also of errors involving opioid narcotics (for example, mor-
improper dose or quantity. phine, fentanyl, hydromorphone) indicate that in many
cases, the health care practitioner may have done one of
Implications and Recommendations the following:
Performance and knowledge deficits and failure to Improperly loaded the PCA syringe/cartridge into the
follow procedures and protocols were the most com- pump, causing air or kinks in the tubing, resulting in lit-
mon causes of administration technique errors report- tle or no drug reaching the patient
ed to MEDMARX. This suggests that health care Incorrectly programmed the PCA pump, delivering
personnel involved in medication administration need either an over- or underdose of the drug
additional education and training. Furthermore, organ- Used the wrong IV tubing to attach the PCA to the
izations should regularly examine the policies and main IV line or improperly connected the PCA pump to
procedures used to guide practitioners in proper drug the patients IV site, causing the tubing to disconnect
administration. Are the policies and procedures in con- The National Coordinating Council for Medication
cert with current best practices? Have staff been ade- Error Reporting and Prevention recommends that users
quately trained on such policies? Is there an of medication administration devices be knowledgeable
assessment of staff competence and understanding on about device function and limitations.7 USP encourages
these policies and procedures? the use of programmable infusion devices with cus-
As illustrated by the case reports, many administra- tomized pump settings that meet institutional guidelines
tion technique errors involve the IV route of administra- for drug dosages for specific patient types and special-
tion. Errors with medications given by this route often ized care areas.6 To prevent staff confusion, organiza-
have the most serious outcomes. USP encourages organ- tions should avoid using different models of infusion
izations to adopt standardized procedures for IV drugs in devices.
all phases of the medication use process,6 including the USP also recommends that infusion pump settings
following elements: should be independently confirmed by two qualified
Prescribers should order only standardized concen- individuals not only upon initiation of high-alert medica-
trations. tions (for example, anticoagulants, cancer chemothera-
Protocols for ordering IV infusions should be clear py agents, insulin, opiates) but for any required dosage
and should eliminate the possibility of calculation errors. adjustment. Documentation of the medication infusion,
Pharmacies should always dispense unit dose rather adjustments, and independent confirmation should be
than multidose containers whenever possible. readily apparent.
Pharmacies should stock or prepare standardized The use of proper and reasonable safety measures
concentrations for all IV medications. can avoid free-flow errors. IV administration cassettes
Solution containers should be labeled with the final that offer antifree-flow mechanisms should be rou-
concentration of the product and the infusion rate as tinely used. Free flow should be avoided when the
appropriate. Appropriate warning labels (for example, IV administration cassette has been removed from

531
September 2005 Volume 31 Number 9
Copyright 2005 Joint Commission on Accreditation of Healthcare Organizations
the pump (for example, during a gown change or to nurses administering medications on the patient care
clear air). unit, although errors in administration technique also
Infusion tubing should be traced from the infusion occur in other phases of medication use and involve
bag to the point of delivery to the patient. If multiple other health care personnel and locations.
infusions and pumps are used on the same patient, each The most commonly reported causes of error have
pump and respective tubing should be readily identifi- been performance deficit, failure to follow procedures or
able and labeled. protocols, and knowledge deficit. Educating and training
Reporting and sharing cases of error events can health care personnel on proper administration tech-
contribute to the collective learning about a drugs niques and use of infusion pumps and dispensing devices
properties (for example, its physical compatibility in could reduce the risk of error. The drug products most
various solutions or its propensity to cause adverse often associated with administration technique errors
effects depending on the rapidity with which it is and patient harm could be targeted in staff education
administered into the body). Subsequently, this learn- and training programs. J
ing can lead to improved and more appropriate admin-
istration techniques.
John P. Santell, M.S., R.Ph., is Director, Educational
Program Initiatives, and Diane D. Cousins, R.Ph., is Vice
Summary President, United States Pharmacopeia (USP) Center for
Wrong administration technique has consistently the Advancement of Patient Safety, Rockville, Maryland.
been one of the most harmful types of medication error Please send correspondence to John Santell, M.S., R.Ph.,
in health systems participating in MEDMARX. jps@usp.org.
Administration technique errors typically are made by

References
1. National Coordinating Council for Medication Error Reporting 2003. Bethesda, MD: American Society of Health-System Pharmacists,
and Prevention: Index for Categorizing Medication Errors. Adopted 2003, pp. 35333535.
July 16, 1996, revised Feb 20, 2001. http://www.nccmerp.org/ 5. McEvoy G.K. (ed.): Phenytoin/phenytoin sodium. In: AHFS Drug
medErrorCatIndex.html (last accessed Jun. 23, 2005). Information 2003. Bethesda, MD: American Society of Health-
2. Joint Commission on Accreditation of Healthcare Organizations: System Pharmacists, 2003, pp. 21122115.
Facts About the Sentinel Event Policy. http://www.jcaho.org/ 6. Hicks R.W., Cousins D.D., Williams R.L.: Summary of information
accredited+organizations/sentinel+event/sefacts.ht. (last accessed submitted to MEDMARX in the year 2002: The Quest for Quality.
Jun. 23, 2005). Rockville, MD: USP Center for the Advancement of Patient Safety, 2003.
3. McEvoy G.K. (ed.): Vancomycin hydrochloride. In: AHFS Drug 7. National Coordinating Council for Medication Error Reporting
Information 2003. Bethesda, MD: American Society of Health-System and Prevention: Recommendations to Reduce Errors Related
Pharmacists, 2003, pp. 470477. to Administration of Drugs. http://www.nccmerp.org/council/
4. McEvoy G.K. (ed).: Phytonadione. In: AHFS Drug Information council1999-06-29.html?USP_Print=true (last accessed Jun. 23, 2005).

532
September 2005 Volume 31 Number 9
Copyright 2005 Joint Commission on Accreditation of Healthcare Organizations

You might also like