Epidemiology of Medication Errors Almut G. Winterstein, PhD

Epidemiology of Medication
Errors
Almut G. Winterstein, PhD
College of Pharmacy, and
Masters of Public Health Program at the College of
Public Health and Health Professions
University of Florida
Epidemiology of Medication Errors
1. Drug therapy outcomes: As good as it gets?
2. Medication errors: How do they look like?
3. In my xxx there are no errors –
What do you know about your own
environment and how do you find out?
“Health Care is a decade or more behind
high-risk industries in its attention to ensure basic
patient safety” (IOM, 2000)
z “... more people die in a given year as a result of
medical errors than from motor vehicle accidents,
breast cancer, or AIDS.”
z “... medication-related error is “one of the most
common types of errors.” [S]ubstantial numbers of
individuals are affected, and it accounts for a
sizeable increase in health care costs.”
Definitions
z Medical error
{Slips and lapses (10-fold overdose, wrong leg
amputated)
{Suboptimal care (not following clinical guidelines or
best practices)
{Errors of omission (indirectly related to medical
intervention: no therapy, underdose)
{Errors of commission (directly related to medical
intervention: overdose, contraindicated therapies)
z Preventable adverse drug event: clinically manifest error
Quality Deficits:
Delays in Implementing Evidence
z Meta-analysis published 1985 emphasized efficacy of
beta-blockade during and after acute myocardial
infarction (AMI) (Yusuf et al. Progr Cardiovasc Dis 1985)
z ISIS-1 Mega-Trial established efficacy in >16,000 MI
cases (ISIS-1, Lancet 1986)
z Study evaluating patient data in 1996:
{ Among the 45,308 AMI patients without contraindications,
50.0% had a beta-blocker as a discharge medication.
(Krumholz et al. JAMA 1998)
Marker for Quality Deficits: Variation
z There was significant variation by state
(30.3% -77.1%).
z Patients with general and family practice physicians as
primary care provider had lower rates
(OR, 0.78; 95% CI, 0.73-0.83).
z Beta-blockers were associated with a 14% lower risk of
mortality at 1 year after discharge.
(Krumholz et al, JAMA 1998)
Life-Cycle of Quality Deficits
% patients with
appropriate care
New
evidence
Clinical
standard
Quality
Deficit
Medical
Error
How many errors do you have?
z At some point in your life have you experienced a medical error in your own
care or that of your family members?
{Yes / No
z In your role as pharmacist have you experienced a medical error that resulted
in serious patient harm in the last year?
{Yes / No
z What is the incidence of medication errors in your institution?
{<0.1%
{<1%
{<5%
{<10%
{>10%
Harvard School of Public Health/Henry J.
Kaiser Family Foundation Survey
(Health & Medicine Week , Feb 2003)
z 35% of physicians reported experiencing a medical
error in their own care or that of a family member at
some point in their life;
z 18% of physicians said an error caused "serious
health consequences" such as death (7%), long-term
disability (6%) or severe pain (11%)
z 29% of all doctors said that in their role as a physician
they had seen a medical error that resulted in serious
harm to a patient in the last year
Evidence Base for the IOM report:
Significance of pADE
z Harvard Medical Practice Study
(Leape, N Engl J Med 1991, 1993)
{ stratified random sample of 51 hospitals in state NY
{ assessment concerned hospital-acquired adverse events
{ adverse event defined as injury that resulted in patients’
death, long-term disability or an extended length of stay in
the hospital
{ primary scope was assessment of cases for litigation and
negligence claims
Harvard Medical Practice Study contd.
z Review of 30,195 charts
z assessment of two core concepts:
{ causality (degree of association between drug and adverse
outcome)
{ preventability
z review of medical charts by two clinical experts, in cases of
disagreement 3rd broke tie
z 0.72% of all patients experienced an adverse drug event
z 0.32% of all admitted patients experienced a preventable
adverse drug event during their hospital stay
Brigham Study Series (Bates et al.)
z Included events classified as “significant, serious, life-threatening, fatal
Author (+)/(-) criteria
Case ascertainment
ADE
(%)
pADE
(%)
Bates
1995
(+) random
Solicited reports & medical
sample of adults of record review, implicit
rd
2 hospitals
judgement by 2 3 broke tie
(-) obstetric units
6.5%
n=4031
1.8%
Bates
1995
(+) all adults
Solicited reports & medical
admitted to 3 units record review; implicit
judgement by 1
6.6%
n=379
1.4%
Bates
1993
(+) all adults
Solicited reports & medical
admitted to 7 units record review, implicit
rd
judgement by 2 3 broke tie
6.4%
n=420
3.6%
Acute Care Estimates of pADEs
z Systematic Review of 10 studies that assessed pADE
incidence and nature (Kanjanarat & Winterstein, AJHP 2002)
{ pADE incidence: median 1.8% (range 1.3-7.8%)
z Cost of pADEs (Bates et al., JAMA 1994)
{ Additional length of stay associated with preventable
adverse drug events: 4.6 days
{ Increase in additional costs: $4685
{ Annual costs for a 700-bed teaching hospital attributable to
preventable ADES $2.8 million
Ambulatory Care Estimates
IOM report:
“..there is evidence indicating that adverse drug
events account for a sizeable number of admissions
to inpatient facilities, but we do not know what
proportion of these are attributable to errors.”
Ambulatory Care Estimates
(Winterstein et al, Ann Pharmacother 2002)
z Meta-analysis of 15 studies reporting prevalence of
preventable drug-related hospital admissions
{ pADEs acquired in ambulatory care leading to
hospitalization
{ pADE prevalence: median 4.3% (range 1.3 – 14%)
{ Stratum-specific prevalence estimates
z Re-admissions vs. all admissions: OR 3.7 (1.5 – 8.9)
z Mean age > 70 vs. ≤ 70y.: OR 2.0 (0.95 – 4.2)
z Omission errors included vs. excluded: OR 1.9 (0.92 – 3.9)
Can you think of medication errors that
frequently cause hospital admission?
Can you think of medication errors that
frequently occur in the ER?
Summary
z Medication errors occur frequently
{ Any instance where new evidence is not implemented can
be considered a medication error
{ Any instance where peers are able to deliver superior
quality than you can be considered medication errors
{ Errors don’t include only causing harm with drugs but also
causing harm by omitting drugs
Patient Safety Solutions
In our opinion, medical error is an
issue for cognitive science and
engineering, not medicine.
(Zhang et al., JAMIA 2002)
Nursing Shortage Poses Serious
Health Care Risk: Joint
Commission Expert Panel Offers
Solutions To National Health
Care Crisis (JCAHO 2002)
Based on overwhelming scientific
evidence, The Leapfrog Group decided to
focus on three practices that have
tremendous potential to save lives by
reducing preventable mistakes in hospitals.
First things first: "medical error" is not
"medicine" it is "error." The discipline
appropriate to its study and diagnosis is not
medicine but theory of error - errorology!
(visualexpert.com)
How do we know?
ƒ What is the most important (prevalent) error cause?
ƒ What is effective in accomplishing the IOM goal for a
50% reduction in medical error by 2005?
ƒ What is efficient?
ƒ What would address the specific concerns of our
institution?
ƒ What would utilize pharmacists’ time and expertise
most optimally?
Description of Medication Errors
z By drug
z By outcome
z By cause
Information on Medication Errors
1. Drug alerts by federal agencies and other organizations / voluntary
reporting
ƒ
ƒ
ISMP, JCAHO
MedMarx, Studies on voluntary reporting systems (Winterstein,
Hatton et al. AJHP 2002)
2. Quality Indicators based on Clinical Guidelines and Best Practices
ƒ
AHRQ, NCQA, JCAHO, CMMS
3. Systematic Studies of medication errors
ƒ
ƒ
Systematic Review of published studies
Solicited institution-specific reporting
ƒ Solicited Medication error reports by an interdisciplinary team
(Winterstein, Johns et al., AJHP 2004)
ƒ Solicited Medication Omission error reports by physician
attendings (Winterstein, Rosenberg et al., SIGGM 2004)
1. Alerts: Institute of Safe Medication
Practice (ISMP)
z Sounds-alike:
New drug Procet® (hydrocodone & acetaminophen)
Similar to: Perocet® (oxycodone & acetaminophen)
z Calculation errors:
Calculators may round results when depending on the number
of decimal places set; in this example the calculator rounded
3.75mg to 4mg
z Look-alike:
ACIPHEX (rabeprazole sodium),
and ARICEPT (donepezil HCl),
MedMARx Database at United States
Pharmacopeia (USP)
z
z
z
z
z
Voluntary database on medication errors
Standardized reporting and classification format
56 participating hospitals in 1999
6,224 medication errors reported
97% of reported medication errors did not cause harm (because
they were averted or not manifest)
z Node: administration (40%), documenting (21%), dispensing
(17%), prescribing (11%), monitoring (1%)
z Most frequent error types: omission (1,689), improper
dose/quantity (1,323), unauthorized drug (751)
z Top causes: performance deficit, protocol not followed,
knowledge deficit
2. CMMS Quality Indicators
(Jencks et al., JAMA 2000)
Quality varies…
(Jencks et al.,
JAMA 2000)
Assessment of Quality across States
(Jencks et al., JAMA 2000 – update 2003)
z Estimates for Florida (ranks 41st)
{ 77% (88%) receive aspirin / 60% (76%) BB within 24 h
{ 78% (90%) receive aspirin / 69% (90%) BB at discharge
{ 29% (49%) of smokers receive smoking cessation counseling
when admitted for MI
{ 76% (89%) admitted with CAP receive Abx within 8h
{ 74% (75%) receive Abx consistent with current
recommendations
{ 69% (82%) of all diabetes patients have annual HbA1c
{ (%) for State with highest ranking (New Hampshire)
Update ranking
3. Systematic Studies
z Prospective follow-up on consecutive admissions or
retrospective chart review
z Expert assessment of
{ causal association between drug therapy and morbidity
{ preventability
z Description by type of drug, type of outcome, type of
drug-related problem
z Limited assessment of causes
pADEs in patients at an intensive care
unit (4.8% pADE incidence) (Darchy et al., 1999)
Adverse outcome
Severe bleeding*
Metabolic disorder
Fluid overload*
Acute renal failure
Respiratory failure
Hyperkalemia
GI bleeding
Cardiac failure
Hypokalemia
% of
ADEs
15.4
10.2
10.2
10.2
10.2
7.7
5.1
5.1
5.1
Medication associated with adverse
event
Oral anticoagulant (& interactive drug )
Miscellaneous
IV fluid infusion
ACEI, cytotoxic drugs
Anesthesia, miscellaneous
ACEI&diuretics, diuretic
NSAIDs
Anesthesia
Diuretic
Systematic review of 10 acute care
studies (Kanjanarat & Winterstein, AJHP 2002)
{ Top drugs: Cardiovascular, psychoactive/CNS,
analgesics, antiinfectives, anticoagulants,
antihyperglycemics
{ Top adverse outcomes: allergic reactions, hepatic or
renal problems, cardiovascular problems, hematologic
problems and bleeding, and central nervous system
problems
{ Top causes: inappropriate prescribing decisions and
inappropriate monitoring
Solicited Report by an Interdisciplinary
(Winterstein,Johns et al, AJHP 2004)
Team
z Eight clinicians on adult medical and surgical, medical hematology
and oncology, bone marrow transplantation, and medical and
cardiac intensive care units of large university-affiliated tertiary
care institution
z 3-month report period in 2002
z Validation through independent review by 2 clinicians
(PharmD/BCPS and MD/internist)
z Real-time assessment of causes
z Validation of reports by 2 independent reviews
z Clustering techniques to find prevalent pattern
Results
{ 240 valid reports (81 discarded): 95 manifest, 94 nonmanifest, 51 averted
{ (240/~5000 admissions with LOS > 1 day = 4.8%)
{ Errors by node
z prescribing (72.5%)
z administration (14.6%)
z dispensing (6.6%)
z transcribing (6.3%)
{ Pharmacy averted 32 errors (18.4%) that were initiated in
the prescribing node, and administering was able to avert
another 10 (7%) that had passed undetected through the
previous nodes.
Top 10 Outcomes
Type of Outcome
Uncontrolled infection
Cardiovascular problems (hyper-/hypotension, arrhythmia)
Renal failure
GI-problems
Neutropenia / thrombocytopenia
Electrolyte imbalances
Oversedation / respiratory depression
Hemorrhage
Uncontrolled pain
Hyperglycemia
N
28
11
9
7
6
5
4
4
4
3
%
30
12
10
7
6
5
4
4
4
3
Top Causes
Type of Cause
N
%
Knowledge deficit (dosing 68, drug selection 23, prophylaxis 3)
94
39
Performance deficit
73
30
Lab values not considered
32
13
Patient condition not considered
12
5
Lack of patient monitoring
9
4
Inappropriate drug administration
8
3
Incorrect drug policy/drug information
6
3
Patient history not considered
6
3
240
100
Total
Top Pattern by Outcome, Node, Drug,
and Drug-related Problem
Pattern of preventable adverse drug events (pADEs)
n
%
Uncontrolled infection, prescribed underdose or failure to prescribe anti-infectives
22
23%
Renal failure, prescribed overdose of anti-infective
4
4%
Respiratory depression, prescribed overdose of CNS drug(s)
4
4%
Uncontrolled pain, prescribed underdose or wrong choice of opioids
4
4%
GI problems, prescribed unnecessary GI drugs
4
4%
Hyperglycemia, prescribed underdose of insulin
2
2%
Hemorrhage, prescribed overdose of blood formation and coagulation drugs
2
2%
Electrolyte imbalance, prescribed underdose of electrolytes
2
2%
Cardiac problems, failure to prescribe electrolytes
2
2%
Uncontrolled infect, failure to transcribe/order anti-infective
2
2%
Total
48
51%
Solicited Report by Senior Attendings
(Winterstein, Rosenberg et al, 2004)
•
•
•
9 attending MDs on general internal medicine services recruited and
trained to report medication errors that:
• Represented errors of omission -- necessary medication not
prescribed, underdosed, or other necessary activities to assure
optimal drug therapy omitted (e.g., monitoring).
• Were clinically manifest / had > 50% likelihood of becoming
manifest. Manifestation defined as death, disability, disease,
therapeutic intervention necessary to prevent further harm.
Medical chart review and provider interview to obtain information on
error etiology and causes.
Validation of reports through investigator consensus
Results
z 34 error reports, 9 discarded
z 25/320 admitted patients with errors (7.8%)
ƒ Lack of drug was more prevalent than underdosing
ƒ Prescribing was associated with more errors (56%) than
administering (36%), transcribing (8%) or dispensing (0%)
ƒ Most prominent error causes were:
ƒ Drug knowledge (e.g., DVT prophylaxis vs. treatment)
ƒ Ineffective communication (e.g., accessibility of MD; explicitness
of PRN orders)
ƒ Performance (provider forgot to order or administer drugs)
ƒ often associated with work flow at rounds, patient transfers,
faulty chart audit procedures (review of orders failed to
detect omitted doses/drugs)
“Highlights”
z Lack of knowledge about heparin dosing for prophylaxis
versus treatment (untreated DVT)
z Failure to obtain or replace off-formulary antipsychotic in
time (uncontrolled psychosis)
z Failure to write complete transfer orders for
bronchodilator (dyspnea, uncontrolled COPD)
z Ineffective communication of morphin PRN order
(uncontrolled pain)
Summary: High-priority Areas of
pADEs
ƒ Preventable hypo- and hyperglycemia
ƒ Preventable acute renal impairment/failure
ƒ Preventable thromboembolic and hemorrhagic events
ƒ Preventable uncontrolled infections
ƒ Preventable respiratory depression and uncontrolled
pain
ƒ Preventable allergic and anaphylactic reactions
What is your top-priority medication
error in the ER and why does it occur?
`