Improving Hospital Medication Safety Using

Information Technology

Tejal Gandhi, MD MPH

Associate Professor of Medicine, Harvard Medical School

Chief Quality and Safety Officer, Partners Healthcare
Handwriting example
 Medication Safety
• The typical hospital medication process has several
steps:
– Ordering- MD orders medication
– Transcribing- nurse copies order onto a paper medication
administration record (MAR)
– Dispensing- pharmacy sends medication to the floor
– Administering- nurse gives medication to patient and
documents this on the MAR

• Medication errors in hospitals are common and can

have serious consequences
– Errors can occur at any stage
 MD

Med Orders

Transcription Pharmacist

Dispensing

Medication Medication on
Admin Record Wards

RN Administration Patient
 Medication Error Frequency and
Potential for Harm
In 10,070 Orders:
530 Medication Errors (1.4 per admission)
35 with potential for harm
5 actually caused harm

• 1 in 100 medication errors results in harm

• 7 in 100 represent potential harm

Bates, DW et al. JGIM 1995

 MD

Med Orders
Ordering
Errors (49%)

Transcription
Transcription Pharmacist
Errors (11%)

Dispensing
Dispensing
Errors (14%)

Medication Medication on
Admin Record Wards

Administration
Errors (26%)
RN Administration Patient
 Main Strategies for Preventing
Errors Using IT
• Tools to improve communication
• Making knowledge more readily accessible
• Requiring key pieces of information
• Assisting with calculations
• Performing checks in real time
• Assisting with monitoring
• Providing decision support
Bates and Gawande, NEJM 2003
 Potential IT Solutions
• Computerized physician order entry (CPOE)
tackles ordering errors
– Computerized writing of orders
– 55% reduction in serious med errors
• Barcoding, electronic medication
administration records (eMAR), and smart
pumps can tackle transcription, dispensing
and administration errors
 Computerized Physician Order
Entry (CPOE)
• Application that allows physicians to write all
orders
– Most things that happen in hospitals occur as a
result of orders
– Computerizing process structures, allows contact
at key times
 Improving the Quality of Drug
Ordering with Order Entry
• Streamline, structure process
– Doses from menus
– Decreased transcription
– Complete orders required
• Give information at the time needed
– Show relevant laboratories
– Guidelines
– Guided dose algorithms
• Perform checks in background
Drug-allergy Dose ceiling Drug-lab
Drug-drug Drug-patient
Allergy to Medication
High Chemotherapy Dose Warning
Impact of CPOE on Medication
Errors

• CPOE reduced medication errors by

80%
• CPOE reduced serious medication
errors by 55%

Bates DW et al. JAMA 1998

Unintended Consequences (UCs)
• Every new technology introduces new errors
• Often poor implementation and poor vendor
design is a significant contributor
• Always a continuous improvement
opportunity
– Brigham and Women’s implemented CPOE in 1994
and is still making improvements
 UCs With Inpatient CPOE
• Workflow
• More work/new work
• Communication
• Overdependence on technology
• Shift in power
• Never ending technology demands
• Emotions
• New errors
Ash, Sittg, Poon et al. JAMIA 2007
 Types of New Errors
• Loss of vigilance
• Loss of visual cues:
– e.g. Writing orders on the wrong chart
• Abuse of technology
– The one-click ‘renew all’
• Propagation of errors, especially with decision
support
• Communication errors
– e.g. During patient transitions
 Overriding of Alerts
• Studies have shown that MDs override clinical
decision support alerts a large percent of the time
– 88% of inpatient DDI alerts overridden (Payne et al. Proc
AMIA 2002)
– 83% of inpatient drug-allergy alerts (Abookire et al. Proc
AMIA 2000)
– 89% of outpatient high severity DDI alerts and 91% of
outpatient drug-allergy alerts (Weingart et al. Arch Intern
Med 2003)
• Overalerting has led to major boycotts of CPOE
systems (e.g. Cedars Sinai)
 Overall Alerting Issues
• Need more studies to maximize
effectiveness of alerts/ minimize over-
alerting

• Issue of how best to display the messages

– Need to learn from other industries (industrial
engineering)
Drug-Pregnancy Level 1
 Potential Strategies to Improve
Alerting
• Creation of streamlined knowledge bases
– Only essential content
– Balance between sensitivity and specificity
• Tiering of alerts is also a possibility
– Hard stop
– Interruptive
– Non-interruptive
• Minimizing interruptions
 Impact of Tiering on Inpatient DDI
Alerts
• Two academic medical centers
• Same knowledge base
• Site A used 3 tiers
• Site B had all of the alerts as interruptive
(Level 2)
• Overall alert acceptance higher at tiered site
(29% vs 10%, p<.001)
Paterno, et al. JAMIA 2009
Tiered Inpatient DDI Acceptance Rates
• Level 1 Acceptance rates
– 100% (hard stop) vs 34% (not a hard stop)
• Level 2 Acceptance rates
– 29% vs 11%
– Likely higher at tiered site since less alert fatigue
because fewer interruptive alerts
 Conclusions
• Streamlined knowledge bases and tiered alerting
have higher acceptance rates
– Especially for very high risk alerts
• What is our ideal acceptance rate??
Sensitivity/specificity? Best way to display?
• More work needs to be done to maximize the clinical
benefits
• Sharing of streamlined knowledge should be
widespread
– No need to reinvent the wheel
 Epidemiology of Dispensing Errors

• Dispensing errors are relatively common in

hospital pharmacies because of the high volume
of medications dispensed
– 44,000 errors/year in a 735-bed hospital (6 million doses/yr)
• Many dispensing errors have potential for harm
– More than 9500 errors with potential to harm patients
occur per year in a 735-bed hospital
– Only 1/3 of these serious errors intercepted prior to
administration
Cina, Gandhi, Churchill, Fanikos, McCrea, Mitton, Rothschild, Featherstone, Keohane, Bates, Poon.
Jt Comm J of Qual & Safety. Jt Comm J of Quality and Safety, Feb 2006
 Pharmacy Barcoding
• Pharmacy technicians use barcode scanning
to verify that the drug they are dispensing
matches the physicians’ orders
 Dispensing Errors and Potential ADEs:
Before and After Barcode Technology Implementation
1.00%
0.88% Before Period (115164
31%
doses observed)
0.80% reduction*
After Period (253984
0.61% doses observed)
0.60% 63%
reduction*
0.40%
0.19%
0.20%
0.07%

0.00%
Dispensing Error Rate Potential ADE Rate

* p<0.0001 (Chi-squared test)

Poon, Cina, Churchill, Featherstone, Rothschild, Keohane, Bates, Gandhi. Annals of Internal Medicine 2006.
Effect of Barcode Technology on Target
Potential ADEs
58% 53%
0.12% 100%
reduction* reduction*
reduction*
0.10%
0.08%
0.08%
0.06%
0.06%

0.04% 0.04%
0.04%
0.03%

0.02%
0.00%
0.00%
Wrong Medication Wrong Wrong Dosage Form
Dose/Strength

Before Period (115164 doses observed)

After Period (253984 doses observed)

* p<0.001 (Chi-squared test)

 Projected Impact at Brigham and
Women’s Hospital
• As we speak, the barcode pharmacy system is
preventing per year:
– >13,500 medication dispensing errors (31%
reduction)
– >6,000 errors with potential for harm (63%
reduction)
 Benefits of Barcode Technology in
the Pharmacy
• Medical costs saved through adverse drug
event reduction, per year
• Increased on-time medication availability on
nursing units
• Improved inventory control
• Formal cost benefit analysis showed break-
even within first year after go-live
– 5-year cumulative net benefit = $3.3M
Maviglia, S et al. Archives of Internal Medicine 2007
 Barcode/eMAR at the Bedside
• Orders flow electronically from CPOE to an
electronic medication administration record (eMAR)
– Eliminates transcription entirely
– Nurses have laptops with eMAR and use this to track what
medications need to be given (administered)
• Nurses use barcode scanning of the medication and
the patient to verify that the drug they are
administering matches the physicians’ orders
– Right drug, right patient, right dose, right time
– eMAR alerts if any of these is incorrect
– Potentially reduces administration errors
 eMAR Hardware
• 2D Imagers
– Both 1 and 2
dimensional bar codes
– Wireless blue tooth
compatible
• Computer Hardware
– Full size laptop
– Complete desktop
functionality
– Mobile carts
Scheduling of medications
Wrong Medication Alert
Wrong Patient Alert
 Impact of Barcode Scanning
Technology on Administration
Errors and Potential Adverse Drug
Events
No Barcode Barcode Relative Reduction
Scanning Scanning (p-value)
(n=6712) (n=7314)
Medication 11.5% 6.7% 41% (p<0.001)
Administration Errors
Potential Adverse Drug 3.1% 1.6% 50.8% (p<0.001)
Events

Poon et al NEJM 2010

37
 Impact of EMar on Nurse Satisfaction

• Pre and post surveys

• Main Results: Nurses feel medication
administration is safer and more efficient after
implementation of barcode technology
Hurley, A et al. Journal of Nursing Administration 2007
 Impact on Nurse Workflow
• 232 2-hour observation sessions before and
after barcode/eMAR implementation

• Primary Result: Proportion of time spent on

medication administration did not change after
barcode/eMAR implementation

• Secondary Result: Proportion of time spent in

presence of patient increased
Keohane C, et al. Journal of Nursing Administration [in press]
 Conclusions
• Barcode technology significantly reduces dispensing,
transcription, and administration errors
• Benefits of the technology outweigh its costs in the
hospital pharmacy
• A well-designed and fully-supported system did not
increase the proportion of time nurses spend on
medication administration
• The technology does not appear to compromise the
amount of time nurses spend with patients.
• Key is involvement of end users from the beginning in
design, hardware selection, and piloting
 IV Medication Safety
• Several studies show that IV medications are
responsible for 54-61% of the most serious
and life threatening potential adverse drug
events.
• Almost all “high risk” drugs (heparin, insulin,
morphine, potassium chloride) are
administered via the IV route.
 Smart Pumps and Medication
Safety
• Barcoding helps ensure right drug, time,
etc
• However, for IV medications, the biggest
error involves programming the infusion
pump
– Manual nursing step
– Barcoding does not address this (yet…)
• Work in progress to automatically program pumps via
wireless communication or barcode scanning
 Features of the “Smart” Pumps
• “Smart” pumps share
safety features of older
pumps
• “Smart” pumps also
equipped with a drug
library
– Provide dose and rate
limits on commonly
used medications
– Provide users with
overdose and
underdose alerts
Case Examples:
Guardrail Near Miss Intercepts
• Dopamine
– entered at 70 mcg/kg/min instead of 7
• Epinephrine
– entered at 32 mcg/min instead of 2
• Heparin
– entered [ ] of 5 units/250 cc rather than 25,000/250
cc
 Summary: Impact on Serious Medication
Errors
MD
Order Entry &
Ordering decision support-
Med Ordering
Errors (49%) 55% reduction

Pharmacist Pharmacy
Transcription
Transcription
Errors (11%) Barcoding-
Dispensing 67% reduction
eMAR- Dispensing
100% reduction Errors (14%)

Medication Medication on -eMAR/barcoding

Admin Record Wards at bedside
- 51% reduction
Administration
Errors (26%)
RN Administration Patient
46
 In Summary
• Non-technology and technology solutions are
both important for improving medication
safety
• Creating a culture of safety and ensuring
action based on events identified is critical
• Technology can provide the high reliability
infrastructure to reduce human error
• Studying the impact of these interventions is
essential