Leigh Briscoe-Dwyer, Pharm.D., BCPS, FASHP

1
Improving Medication Safety in the ICU

• Leigh Briscoe-Dwyer, Pharm.D., BCPS, FASHP
• Chief Pharmacy and
• Medication Safety Officer
• North Shore LIJ Health System
• September 19, 2015

2
CONFLICT OF INTEREST DISCLOSURE

• I have no conflicts to report

3
Objectives

• 1. Discuss national regulations and local (but
  universally adopted) practices related to
  medication safety and adverse drug reaction
  reporting and high reliability organizations.
• Examine the use of clinical decision support
  systems (CDSS) in drug-drug interaction (DDIs)
  identification and prevention.
• Describe off-label medication use in the ICU
  setting, implications for patient safety, and
  strategies to minimize patient harm when
  medications are being used off-label.
• Describe the use of trigger tools and the
  optimization of technology to improve medication
  use processes.

4
Medication Safety Organizations

• Agency for Health Care Research and Quality
• American Association of Critical-Care Nurses
  (AACN)
• American College of Physicians (ACP)
• American Hospital Association
• American Nurses Association (ANA)
• American Organization of Nurse Executives (AONE)
• Anesthesia Patient Safety Foundation (APSF)
• Association of periOperative Registered Nurses
  (AORN)
• American Society for Healthcare Risk Management
• American Society of Health-System Pharmacists
• Annenberg Center
• Centers for Disease Control and Prevention (CDC)

• ECRI
• Institute for Healthcare Improvement (IHI)
• Institute of Medicine
• Institute for Safe Medication Practices (ISMP)
• The Joint Commission
• KLAS
• Leapfrog
• National Coordinating Council for Medication
  Error Reporting and Prevention
• National Patient Safety Foundation
• National Quality Forum (NQF)
• Society of Critical Care Medicine (SCCM)
• US Food Drug Administration
• VA - National Center for Patient Safety

5
Why Do We Still Have Problems with Medication
Safety?

• 10,000 prescription medications exist
• Nearly 1/3 of adults in the US take 5 or more
  medications
• ADEs account for 700,00 ED visits and 100,000
  hospitalizations annually
• 5 of hospitalized patients affected

AHRQ PSNet Patient Safety Primer
http//psnet.ahrq.gov
6
Question

• 1. Which of the following best describes the
  relative incidence of adverse drug events (ADE),
  adverse drug reactions (ADR), and medication
  errors (ME)?
• ME gt ADR gt ADE
• ME gt ADE gt ADR
• ADE gt ADR gt ME
• ADR gt ME gt ADE

7
Question

• 1. Which of the following best describes the
  relative incidence of adverse drug events (ADE),
  adverse drug reactions (ADR), and medication
  errors (ME)?
• ME gt ADR gt ADE
• ME gt ADE gt ADR
• ADE gt ADR gt ME
• ADR gt ME gt ADE

8
Detection and Classification of Adverse Drug
Events
Morimoto T et al. Qual Saf Health Care. 2004
13306-14
9
Nomenclature

• Adverse Drug Reaction (ADR) - A response to a
  medicinal product that is noxious and unintended
  and that occurs at doses normally used in humans
  for the prophylaxis, diagnosis, or treatment of
  disease
• Adverse Event (AE) - An injury, large or small,
  caused by the use (including non-use) of a drug,
  test, or medical treatment. This may be as
  harmless as a drug rash or as serious as death.
  (modified from IHI definition of an adverse drug
  event or ADE.)

10
Medication Error Categorization
11

• What do you call an organization/industry that is
  complex and riskyBut very safe?
• Highly Reliable Organization

12
What is a Highly Reliable Medication Use Process?

• The measurable capability of a process to perform
  its intended function in the required time under
  commonly occurring conditions
• Ask 5 people how they perform a task or a
  process, if they all are the same your process
  is reliable
• Action Item
• Choose a timely topic on mediation use
• Survey people involved in the process
• Evaluate for reliability

13
Different Views of Reliability1 million doses
dispensed/month
Reliability Unreliability Sigmas (approximate)
0.9 10-1 1
0.99 10-2 2
0.999 10-3 3
0.9999 10-4 4
0.99999 10-5 5
0.999999 10-6 6
ERRORS 12,000/year 1000/month 33/day
14
Design for Reliability

• Level 1. Intent, vigilance and hard work
• STANDARDIZATION
• Level 2. Design informed by reliability science
  and research in human factors
• Level 3. Design of high reliability
  organizations
• (Weick)

15
Trigger Tools

• Institute for Healthcare Improvement
• Small percentage of adverse events (AE)
  voluntarily reported
• Use of "triggers," or clues, to identify AE in an
  organization captures larger percentage
• Reporting increased by 10 times when a Global
  Trigger Tool was utilized

Classen DC et al. Journal of Patient Safety. 2008
Sep 4(3)169-177. Adler L et al. Journal of
Patient Safety. 2008 Dec 4(4)245-9. Classen DE
et al. Health Aff. 2011 April 30(4)581-9.
16
Use of Global Trigger Tools
Classen DE et al. Health Aff. 2011 April
30(4)581-9.
17
50 Dextrose as a Trigger Tool
Site Doses Dispensed
CCMC 9
FHH 699
FRK 740
LIJ 2433
NSUH 1849
PLV 205
SSH 462
SYO 40
ZHH 1
Total 6438
18

• Drug Drug Interactions and Clinical Decision
  Support

19
The promise and problem with alerts

• To err is human. The EMR can be a wise
  detail-oriented consultant for certain go/no-go
  decisions -- Dear Dr., this patient had an
  anaphylactic reaction to Bactrim. Please
  consider prescribing an alternative antibiotic.
  
• If the EMR commonly alerts in an unwise manner,
  providers may fail to attend to future alerts.

20
How do we Treat Alerts?

• Drug-drug interaction alerts
• One month study of 39,893 alerts with 45,983
  orders (0.87 alerts per med order).
• 1,176/45,983 (2.6) of med orders had a
  pharmacist intervention.
• 113/45,983 (0.2) of med orders had a pharmacist
  intervention involving the alert.
• Only 113/39,893 (0.03) of drug-drug interaction
  alerts on med orders resulted in a pharmacist
  intervention.
• Conclusion The interventions made by
  pharmacists were not a result of the alerts that
  fired.

21
How often does the boy EMR cry wolf for
medication orders?
22
Duplicate alerts what triggers an alert?
Exact match or two drugs in the same class from
the following list

• Corticosteroids
• H2 blockers
• NSAIDs
• Oral anti-diabetic agents
• Parenteral anticoagulants
• Parenteral opiates
• Penicillins
• Phenytoins
• Potassium containing products
• PPIs
• Quinolones
• SSRIs
• Statins

• ACE Inhibitors
• Acetaminophen containing products
• Alpha-1 blockers
• Aminoglycosides
• Anticoagulants/Antiplatelets
• Benzodiazepines
• Beta-blockers
• Calcium channel blockers
• Carbapenems
• Cephalosporins
• Combination analgesics

23
Duplicate alerts

• 40 of these alerts are for drug class
• 98 of the time, the provider finalized the
  order.
• 75 of the time, the provider did not explicate
  the reason for overriding the alert. When they
  acknowledged with text
• provider aware and approved
• to be addressed by primary provider
• OK
• 96 of the duplicate orders that were finalized
  by providers were verified by pharmacy, 69 of
  which were verified within 10 minutes of the
  order being authored, suggesting that no
  conversation with the provider took place.

24
The Joint Commission Leadership in Healthcare
Institutions (2009)

• Develop a graduated system of safety alerts in
  the new technology that helps clinicians
  determine urgency and relevancy.
• Consider skipped or rejected alerts as important
  insight into clinical practice.
• Decide which alerts need to be hard stops when
  using the technology and provide appropriate
  supporting documentation.

25
Recommendations

• Establish a governance group responsible for
  managing changes in alerts.
• Change the mindset regarding alerts from a 10
  commandments view to something we are
  responsible for managing.
• Recognize that the primary purpose of the
  governance group is to improve patient safety by
  driving down low-value alerts.
• Get rid of duplicate therapy alerts for PRNs
  that are not true duplicates.

26
Off Label Drug use in the ICU

• FDA is responsible for review and approval of
  medications for specific indications for which
  there is safety and efficacy data
• Once a medication is approved, there is usually
  no limit imposed on FDA for its use
• Off label use is often the standard of care
• Oncology, pediatrics, critically ill patients

27
Off Label Use in Surgical ICU

• 465 orders for 80 medications were reviewed
• 26.5 of orders were described as off-label
• Indications
• Infections, stress ulcer prophylaxis, seizure
  prophylaxis and treatment
• Reasons for use
• Unapproved indications (66)
• Dosing schedule (27)
• Method of administration (17)

Albaladejo P, Caillet B, Moine P et al. In
French Presse Med 2001301484-1488
28
Off Label Medication Use in Adult Critical Care
Patients

• Prospective observational study of 37 ICUs from
  24 sites over a 24 hour period
• 414 patients enrolled
• 5237 medication orders
• 1897 orders were off-label (36.2)

Lat I, Micek S, Janzen J et al. Journal of
Critical Care 2011 2689-94
29
Off Label Medication Use in Adult Critical Care
Patients

• Most frequent drug classes
• Bronchorespiratory
• Gastrointestinal
• Immunology
• 89 of off label orders were written after
  patients were admitted to ICU
• 928 (48.3) of off label use had grade C or no
  evidence to support such use

30
Going Off Label Without Going Off Course

• Which characteristics of off label use require
  greater scrutiny?
• Can we differentiate off label use based on
  available evidence?

Largent E, Miller F, Pearson S. Arch Int Med
2009. 169 (19) 1745-1747.
31
Signals for Scrutiny

• New drugs
• Rule of Thumb 3 5 years of observational study
  before making a judgment of risk
• Novel off label use
• Combination therapies
• Drugs with known serious adverse effects
• High-Cost drugs

32
Levels of Evidence to Guide Off-label
Prescribing

• Supported
• Suppositional
• Investigational

33
Supported Off Label Use

• Moderate to high level of certainty in net health
  benefit
• Prescribe in same manner as when indication
  exists
• Discuss with patient/family benefits and risks of
  proposed treatment

34
Suppositional Off Label Use

• Low level of certainty in net health benefit
• Is it better than no treatment?
• Consultation and Second Opinion recommended
• Risks and benefits should be clearly communicated
  to patients and documented in patients chart
• Collect data on outcomes

35
Investigational Off Label Use

• Very low level of certainty in net health benefit
• Should be limited to context of research
  protocols
• Informed Consent

36

• Evidence and Extrapolation
• Mechanisms for Regulating off-Label Uses of Drugs
  and Devices

37
Evidence and Extrapolation

• Diagnosis Extrapolation
• Using an existing drug to treat a new condition
• Using quetiapine to treat anxiety rather than
  schizophrenia
• Patient Extrapolation
• Using an existing drug to treat a new population
  with a given condition
• Treating children rather than adults

Abbot R and Ayres I. Duke Law Journal 2014.
643 377- 435
38
Evidence and Extrapolation

• Dosage Extrapolation
• Using an existing drug for a new duration or
  schedule
• Using a drug indefinitely for schizophrenia when
  studies have looked at 6 weeks
• Treatment Extrapolation
• Using an new drug that is related to an approved
  counterpart
• Using extended release quetiapine based on
  evidence that immediate release is safe and
  effective

Abbot R and Ayres I. Duke Law Journal 2014.
643 377- 435
39
Evidence and Extrapolation Recommendations

• Improve level of reporting of off label use
• Expand post-market testing in the area of
  off-label use
• Create a tiered labeling system to allow for
  distinctions for payers and litigation
• Red Box Warning to prohibit off label use
• Gray Box Warning to block reimbursement

40
optimization of technology to improve medication
use processes

• NSHS has a single enterprise drug library for IV
  Infusion pumps
• AKA Smart Pumps
• Technology must be reviewed and maintained
  continually to get the maximum benefit

41
Propofol

• ICU sedation in intubated mechanically-ventilated
  patients
• Continuous infusion
• Initial 5 mcG/kG/minute
• Increase by 5 to 10 mcG/kG/minute every 5 to 10
  minutes until desired sedation level is achieved
• Usual maintenance 5 to 50 mcG/kG/minute 
• NS-LIJ Critical Care Library Guardrails
• Soft maximum of 75 mcG/kG/min

42
Propofol Medication Use Evaluation (MUE)

• LIJMC Medical Intensive Care Unit (MICU)
• 23 patients from April through June 2015
• Procedure
• Ran a report using SCM of all patients currently
  with an active order for propofol infusion
• Obtained the Serial Numbers on the pump
• Ran a report using Guardrails data to look at
  alerts that had fired
• Goals
• Evaluate whether data documented in SCM matched
  the data we obtained from the pumps
• In the case that an alert fired - assess any
  adverse events

43
Propofol MUE

• Data we evaluated
• Accuracy of the patient weight entered in the
  pump
• Rate of infusion programmed when outside soft
  maximum of 75 mcG/kG/min
• Pump is being used appropriately for continuous
  infusions and not for bolus dosing
• Appropriate documentation is found in SCM (e.g.
  order placed, flowsheets, eMAR)
• Potential Adverse Effects
• HR, BP, concomitant use of vasopressors

44
Results Infusion Rate Alerts Overridden

• 4 instances identified when the rate entered
  exceeded soft maximum of 75 mcG/kG/min
• Alerts were overridden

Alaris Pump Data Alaris Pump Data SCM Data SCM Data SCM Data
Programmed Dose patient received Order in SCM Documentation in eMAR or flowsheet Adverse Effects Identified
999 mL/hr (3330 mcG/kG/min) x 5 mL 50 mG 20 mG IVP x 2 40 mG eMAR 20 mg IVP x 2 40 mG None
999 mL/hr (3165 mcG/kG/min)x 5 mL 50 mG No order for IV bolus or increase in rate of infusion None None
999 mL/hr (3165 mcG/kG/min) x 2 mL 20 mG No order for IV bolus or increase in rate of infusion None None
a. 999 mL/hr (1276 mcG/kG/min)x 76 mL b. Reprogrammed to 30 mL/hr 767 mG 87 mG excess No order for IV bolus or increase in rate of infusion None None
45
Infusion Rate Alerts Reprogrammed

• 2 instances identified where the use of
  Guardrails prevented potential errors and
  adverse events
• Alerts fired resulted in reprogramming to
• 75 mcG/kG/min

Reprogramming Due to Guardrails Reprogramming Due to Guardrails
Rate initially entered Rate entered after alert fired
38 mL/hr (100 mcG/kG/min) 20 mL/hr (50 mcG/kG/min)
37 mL/hr (85 mcG/kG/min) 17 mL/hr (40 mcG/kG/min)
46
Weight Change Alerts

• Importance of utilizing weight documented in SCM

Issues Identified Entering Patient Weights Issues Identified Entering Patient Weights Issues Identified Entering Patient Weights Issues Identified Entering Patient Weights Issues Identified Entering Patient Weights Issues Identified Entering Patient Weights
Patient 1 Patient 2 Patient 3 Patient 4 Patient 5
Weight entered 66.4 kG 50 kG 106 kG 70 kG 70 kG
Alert from Pump 64.4 kG 51.3 kG 106.3 kG 63.8 kG 61.8 kG
Actual weight documented in SCM 64.4 kG 51.3 kG 106.3 101.3 kG 68.1 kG
Action taken 64.4 kG 51.3 kG 106.3 63.8 kG 61.8 kG
47
MUE RESults

• Presented at System P and T
• Bolus dosing vs. continuous infusion
• Area for medication errors NEVER EVENT
• Appropriate documentation in SCM (order placed,
  rates in flowsheets)
• Importance of entering the correct weight that is
  documented in SCM

48

• Questions ?