Transcript Designing An Effective Product Recall Program

Using Risk Information in Product
Recall Determinations
Edward J. Heiden, PhD
President
Heiden Associates, Inc.
International Bar Association
2006 Conference
September 18, 2006
Core Areas of Expertise

Product Safety and
Risk Analysis
– Risk and hazard
pattern evaluation
– Product population/
incident projections
– Comparative risk
assessment
– Brand-specific fire risk
analysis
Baseball Injuries
Collision w/
player
Hit with ball
6%
15%
Fell/ sliding
25%
Hit with bat
29%
Other injury
pattern
25%
Core Areas of Expertise

Product Recalls
– Program design and
management support
– Outreach analysis and
options
– Effectiveness
evaluations
– Program audit and
close-out
Possible Corrective Actions





Changing product design
Withdrawing products from the distribution chain
Sending information and warnings about use of
products to consumers
Modifying products at customers’ premises or
elsewhere
Recalling products from consumers for
replacement or refund
Risk-Based Decisions about
Corrective Actions

Identify hazards with actual or potentially
serious risks and respond quickly
 Avoid recalls of products with marginal or
non-existent risks of harm
Need to use all available data to make
informed, defensible decisions
Data Requirements
for Recall Determinations

Representative of product sales, use and
incident profiles
 Documented incident data collection and
reporting procedures
 Adequate detail to identify specific types of
products or hazards
 Few sources meet all 3 tests!
Useful Types of Data

Company/Proprietary
– Warranty returns, consumer complaints
– Claims, lawsuits
– Company/consultant testing

Public Databases
– Product-related Injuries, fatalities
– Fire, poisoning incidents
– Accident investigations, special studies
Public Data Available (CPSC)

National Electronic Injury
Surveillance System
(NEISS)
– Statistically
representative sample
– Demographic and
injury-characteristics
– Brief narrative
descriptions of
incidents
Sample Narrative:
“Pressure cooker opened.
Diagnosis: shoulder burn”
Public Data Available (CPSC)

Other Databases:
Accident
Investigations,
Incident Reports
– Not statistically
representative
– Varied sources of
referral/reporting
– More detailed incident
descriptions
Sample Narrative:
While a brand new pressure cooker
was in use on an electric range at
the consumer's home, the lid to it
blew-off & struck the light & hood
above the range, causing the light
and its cover to shatter. While the
37 year old wife was cleaning up
the debris, she cut the middle
finger to her right hand on some
broken glass from the light bulb.
Other U.S. Public Data Sources





Fire Incidents (NFIRS)
Workplace Injuries (BLS, OSHA)
Transportation Accidents (FARS, GES)
Poisoning/Chemical Ingestion (TESS)
Medical Device Problems (MDRs)
 Varied levels of statistical reliability and detail
Injury Database (IDB) - EU

EHLASS (European Home and Leisure Accident
Surveillance System)
–
–
–
–
–
–
–
–
Place of Occurrence
Time of Injury
Mechanism of Injury
Activity
Type of Sports
Type of Injury
Product causing the Injury
Accident description
Next Steps for the IDB

Extend data collection to include road transport,
workplace and violence
 Gain more information on products and services
 Involve new Member States in injury surveillance
 Develop statistical procedures for the estimation
of population-based injury rates
How Best to Use these Data?
 Precautionary
Principle
 Structured Risk Analysis
 Baseline/Comparative Risk
Assessment
1) Precautionary Principle
If the potential consequences of an
action are severe or irreversible, in the
absence of full scientific certainty the
burden of proof falls on those who
would advocate taking the action.
Application to Recalls

“Scientific proof” means appropriate
economic/statistical methodology

“Action” means “decision not to recall a
product” with some potential hazard
The EU GPSD

A safe product "does not present any risk or
only the minimum risks compatible with the
product's use, considered to be acceptable
and consistent with a high level of
protection for the safety and health of
persons."
Benefits of Using the Principle

Effectiveness of recall diminishes if product
is in consumers’ hands for long time
 Addresses serious or frequent hazards that
might not be apparent at the time of the
determination
 Reduces potential for criticism due to
inaction
Costs of Using the Principle

Source of hazard is not always well understood, so
the remedial action must be validated before
initiating the recall
 Misdirects resources to hazards that do not
represent meaningful risks to consumers
 Causes recall notification “clutter” that makes it
harder to reach consumers when there is a serious
problem
2) Structured Risk Analysis

Attempt to develop qualitative, and if at all
possible, quantitative measures of the risk
associated with consumer use of the product

Address specific criteria, such at those set
forth in
– Consumer Product Safety Act (US)
– GPSD Risk Assessment Procedure (EU)
CPSA Factors

Pattern of Defect
 Number in Commerce
 Severity of Risk
 Likelihood of Injury
 No formal specification of how factors contribute
to overall recall decision
GPSD Framework

Evaluate Overall Gravity of Outcome
– Probability of hazard
– Severity of potential damage

Exposure of Vulnerable Populations
– Children
– Elderly

Adequacy of Warnings/Obviousness of
Hazard
RAPEX: Probability and Severity
RAPEX: Risk  Action Required
Application: Yo-Yo Balls




Potential suffocation hazard if
wrapped around child’s neck
186 incidents, 11-15 million
balls made, no actual
suffocation incidents
US: CPSC issues warning and
advises to exercise caution
Canada: immediate prohibition
on advertising, sale or
importation
Benefits of Statistical Analysis
Meets need for “scientific determination”
 Responds to regulatory requirements
 Provides verifiable basis for making a
decision
 Allows for revision as new data become
available

Problems with Statistical Analysis
No single way to do it “right” – can get
different results from same data
 Relevant information may not be available,
may be of variable quality, or may be
difficult to interpret
 Lack of familiarity with data sources and
methods – particularly their strengths and
weaknesses

3) Baseline / Comparative
Risk Assessment

Determine whether hazard represents a
meaningful increment of the ordinary
(“baseline”) risk associated with use of the
product
 Compare risks on a usage- or exposureadjusted basis with those from comparable
products and/or activities
Application: Electric Shock
Hazards

Assignment:
– Determine
whether a
toaster with a
low potential
for shock
incidents
should be
recalled

Assessment Conclusions:
– Risk of shock was low
compared to overall level of
shock risk from using
toasters
– Risk of shock from using
toasters and other small
kitchen appliances is low
compared to many other
common household
electrical products
# of ER-Treated Shock Injuries
Requiring Hospitalization
Avg./
Product Group
Year
Share
124
24%
Home Structures & Construction Materials
82
16%
Home Workshop Apparatus, Tools & Attachments
60
12%
General Household Appliances
50
10%
Yard & Garden Equipment
46
9%
Home Communication, Entertainment & Hobby Equip.
40
8%
0
0%
Home Furnishings & Fixtures
Small Kitchen Appliances
Benefits of Comparative Risk
Assessment

Shares benefits of structured risk analysis
 Focuses attention and effort on meaningful
risks
 Places risks in the context of consumer
expectations and experience with similar
types of products
Costs of Comparative Risk
Assessment

Responsibility to ensure that “apples to
apples” comparisons are made:
– product comparability
– proper adjustment for sales/use
– nature and severity of injuries taken into account

In some cases, gaining regulatory acceptance
of analysis may be more difficult
Looking Forward

Data available to inform recall decisions is
improving and becoming more readily
available
 Expanding acceptance of, and more
demanding expectations for, analysis used
to make risk-based recall determinations