This is the accessible text file for GAO report number GAO-11-468 
entitled 'Medical Devices: FDA Should Enhance Its Oversight of 
Recalls' which was released on June 22, 2011. 

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as 
part of a longer term project to improve GAO products' accessibility. 
Every attempt has been made to maintain the structural and data 
integrity of the original printed product. Accessibility features, 
such as text descriptions of tables, consecutively numbered footnotes 
placed at the end of the file, and the text of agency comment letters, 
are provided but may not exactly duplicate the presentation or format 
of the printed version. The portable document format (PDF) file is an 
exact electronic replica of the printed version. We welcome your 
feedback. Please E-mail your comments regarding the contents or 
accessibility features of this document to Webmaster@gao.gov. 

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately. 

United States Government Accountability Office: 
GAO: 

Report to Congressional Requesters: 

June 2011: 

Medical Devices: 

FDA Should Enhance Its Oversight of Recalls: 

GAO-11-468: 

GAO Highlights: 

Highlights of GAO-11-468, a report to congressional requesters. 

Why GAO Did This Study: 

Recalls are an important tool to mitigate serious health consequences 
associated with defective or unsafe medical devices. Typically, a 
recall is voluntarily initiated by the firm that manufactured the 
device. The Food and Drug Administration (FDA), an agency within the 
Department of Health and Human Services (HHS), oversees implementation 
of the recall. FDA classifies recalls based on health risks of using 
the recalled device—-class I recalls present the highest risk 
(including death), followed by class II and class III. FDA also 
determines whether a firm has effectively implemented a recall, and 
when a recall can be terminated. This report identifies (1) the 
numbers and characteristics of medical device recalls and FDA’s use of 
this information to aid its oversight, and (2) the extent to which the 
process ensures the effective implementation and termination of the 
highest-risk recalls. GAO interviewed FDA officials and examined 
information on medical device recalls initiated and reported from 2005 
through 2009, and reviewed FDA’s documentation for a sample of 53 (40 
percent) of class I recalls initiated during this period. 

What GAO Found: 

From 2005 through 2009, firms initiated 3,510 medical device recalls, 
an average of just over 700 per year. FDA classified the vast 
majority—-nearly 83 percent-—as class II, meaning use of these 
recalled devices carried a moderate health risk, or that the 
probability of serious adverse health consequences was remote. Just 
over 40 percent of the recalls involved cardiovascular, radiological, 
or orthopedic devices. FDA has used recall data to monitor individual 
recalls and target firms for inspections. However, it has not 
routinely analyzed recall data to determine whether there are systemic 
problems underlying trends in device recalls. Thus, FDA is missing an 
opportunity to use recall data to proactively identify and address the 
risks presented by unsafe devices. 

Figure: Number of Medical Device Recalls Initiated from 2005 through 
2009: 

[Refer to PDF for image: stacked vertical bar graph] 

Recall initiation year: 2005; 
Class I recall (highest risk): 30; 
Class II recall (moderate risk): 133; 
Class III recall (lowest risk): 503; 
Total recalls: 666. 

Recall initiation year: 2006; 
Class I recall (highest risk): 18; 
Class II recall (moderate risk): 110; 
Class III recall (lowest risk): 530; 
Total recalls: 658. 

Recall initiation year: 2007; 
Class I recall (highest risk): 25; 
Class II recall (moderate risk): 91; 
Class III recall (lowest risk): 612; 
Total recalls: 728. 

Recall initiation year: 2008; 
Class I recall (highest risk): 17; 
Class II recall (moderate risk): 89; 
Class III recall (lowest risk): 690; 
Total recalls: 796. 

Recall initiation year: 2009; 
Class I recall (highest risk): 41; 
Class II recall (moderate risk): 57; 
Class III recall (lowest risk): 564; 
Total recalls: 662. 

Source: GAO analysis of FDA data. 

[End of figure] 

Several gaps in the medical device recall process limited firms’ and 
FDA’s abilities to ensure that the highest-risk recalls were 
implemented in an effective and timely manner. For many high-risk 
recalls, firms faced challenges, such as locating specific devices or 
device users, and thus could not correct or remove all devices. FDA’s 
procedures for overseeing recalls are unclear. As a result, FDA 
officials examining similar situations sometimes reached opposite 
conclusions on whether recalls were effective. FDA had also not 
established criteria, based on the nature or type of devices, for 
assessing whether firms corrected or removed a sufficient number of 
recalled devices. Additionally, FDA’s decisions to terminate completed 
recalls—-that is, assess whether firms had taken sufficient actions to 
prevent a recurrence of the problems that led to the recalls-—were 
frequently not made within its prescribed time frames. Finally, FDA 
did not document its justification for terminating recalls. If 
unaddressed by FDA, the combined effect of these gaps may increase the 
risk that unsafe medical devices could remain on the market. 

What GAO Recommends: 

To aid its oversight of the medical device recall process, FDA should 
routinely assess information on device recalls, develop enhanced 
procedures and criteria for assessing the effectiveness of recalls, 
and document the agency’s basis for terminating individual recalls. 
HHS agreed with GAO’s recommendations. 

View [hyperlink, http://www.gao.gov/products/GAO-11-468] or key 
components. For more information, contact Marcia Crosse at (202) 512-
7114 or crossem@gao.gov. 

[End of section] 

Contents: 

Letter: 

Background: 

Firms Initiated Several Thousand Medical Device Recalls, but FDA Has 
Not Routinely Analyzed This Information to Aid Its Oversight of 
Recalls: 

Gaps in the Medical Device Recall Process Limit the Effective 
Implementation and Termination of the Highest-Risk Recalls: 

Conclusions: 

Recommendations for Executive Action: 

Agency Comments: 

Appendix I: Tools to Help Identify Recalled Devices and Planned 
Initiatives for Assisting with Device Recalls: 

Appendix II: Class I Medical Device Recalls from Our Sample for Which 
Firms Were Unable to Correct or Remove All Devices: 

Appendix III: Comments from the Department of Health and Human 
Services: 

Appendix IV: GAO Contact and Staff Acknowledgments: 

Tables: 

Table 1: Comparison of FDA Medical Device Classification and Recall 
Classification Levels, by Risk: 

Table 2: Number and Percentage of Medical Device Recalls Initiated 
from 2005 through 2009, by Medical Specialty Area: 

Table 3: Number and Percentage of Medical Device Recalls Initiated 
During 2008 and 2009, by Root Cause: 

Table 4: Number, Percentage, and Timeliness of Correction or Removal 
Reports Submitted for a Sample of Class I Medical Device Recalls 
Initiated from 2005 through 2009, by FDA District: 

Table 5: Examples of Available Recall Alert Services: 

Table 6: Class I Medical Device Recalls for Which Firms Were Unable to 
Correct or Remove All Devices: 

Figures: 

Figure 1: The Voluntary Medical Device Recall Process Followed by FDA, 
Firms, Customers, and Device Users: 

Figure 2: Number of Medical Device Recalls Initiated from 2005 through 
2009: 

Figure 3: Status of Medical Device Recalls Initiated from 2005 through 
2009: 

Figure 4: Average Time from Initiation to Termination for Medical 
Device Recalls Initiated from 2005 through 2009: 

Figure 5: Percentage of Medical Device Recalls Terminated More Than 90 
Days after Completion Date, for Recalls Initiated in Calendar Years 
2005 through 2009: 

Figure 6: Percentage of Our Sample of Class I Medical Device Recalls 
Initiated from 2005 through 2009 for Which Recall-Related Inspections 
Were Conducted, by FDA District Office: 

Figure 7: Status of Our Sample of 53 Class I Medical Device Recalls 
Initiated from 2005 through 2009: 

Figure 8: Example of a Potential Unique Device Identification Label: 

Figure 9: Timeline of Events for the Development of the Unique Device 
Identification Initiative: 

Abbreviations: 

CDRH: Center for Devices and Radiological Health: 

FDA: Food and Drug Administration: 

FDAAA: Food and Drug Administration Amendments Act of 2007: 

HHS: Department of Health and Human Services: 

ORA: Office of Regulatory Affairs: 

PMA: premarket approval: 

RES: Recall Enterprise System: 

UDI: unique device identification: 

[End of section] 

United States Government Accountability Office: 
Washington, DC 20548: 

June 14, 2011: 

The Honorable Charles E. Grassley: 
Ranking Member: 
Committee on the Judiciary: 
United States Senate: 

The Honorable Herb Kohl: 
Chairman: 
Special Committee on Aging: 
United States Senate: 

Each day, millions of individual medical devices produced by thousands 
of manufacturing establishments in the United States and overseas are 
used in hospitals, physicians' offices, and other health care settings 
to diagnose, treat, or prevent illness. For example, in 2007 medical 
devices were involved in 45 million inpatient procedures. Also, there 
were approximately 117 million visits to hospital emergency 
departments, 89 million hospital outpatient visits, and 994 million 
visits to physicians' offices, which likely all involved the use of 
one or more medical devices. Medical devices include those that 
present little risk--such as tongue depressors and elastic bandages--
and those that are used specifically to sustain or support life--such 
as pacemakers and artificial heart valves. Medical devices are an 
integral part of patient care. If one proves to be defective or unsafe 
once it is in widespread use, the ramifications can be severe, 
potentially resulting in permanent injuries or deaths to patients or 
providers using the device. 

A recall is an important remedial action that can mitigate the risk of 
serious health consequences associated with a defective or unsafe 
medical device. Generally, the firm that manufactured the device 
voluntarily initiates a recall after it has discovered a problem based 
on its assessment of complaints or reports of safety issues it has 
received. The Food and Drug Administration (FDA), an agency within the 
Department of Health and Human Services (HHS), is responsible for 
ensuring that medical products sold in the United States are safe and 
effective. In some cases, FDA identifies a problem with a device based 
on its own oversight, such as an inspection of an establishment where 
a device is manufactured. In response, the firm may voluntarily 
initiate a recall. Additionally, FDA has the authority to require that 
the firm initiate a device recall when there is a serious risk to 
public health and the firm has not done so.[Footnote 1] 

FDA's role in the voluntary recall process is generally to oversee a 
firm's management of a recall, which includes monitoring the progress 
of the recall. FDA has issued guidance to aid firms in conducting such 
recalls and has also established internal procedures to govern its 
oversight of the recall process.[Footnote 2] In addition, FDA assigns 
each recall a classification level--high, moderate, or low--based on 
its assessment of the degree of risk that is posed by the continued 
use of the device. An effective and timely recall depends on actions 
taken by the firm manufacturing the device, companies distributing the 
device, users of the device--such as hospitals, physicians, and 
patients--and FDA. Recalling firms are responsible for alerting FDA, 
their distributors, and users of the device about the recall. The 
recalling firm provides instructions on steps to be taken to fix the 
device or advises parties to discontinue its use. These parties must 
follow the firm's instructions in order to effectively implement the 
recall. 

Despite efforts by recalling firms, FDA, and others, there have been 
reported incidents where individuals were seriously injured or died 
due to continued use of defective devices that had been recalled. You 
expressed concern with the effectiveness of the medical device recall 
process and asked us to follow up on our 1998 report on FDA's 
oversight of medical device recalls.[Footnote 3] Our preliminary 
findings were included in an April 2011 hearing on the reform of the 
medical device approval process before the Senate Special Committee on 
Aging.[Footnote 4] This report identifies (1) the numbers and 
characteristics of medical device recalls initiated from January 1, 
2005, through December 31, 2009, and the extent to which FDA uses this 
information to aid in its oversight of recalls, and (2) the extent to 
which the medical device recall process ensures the effective 
implementation and termination of those classified as high-risk 
recalls. 

To identify the numbers and characteristics of voluntary medical 
device recalls, we obtained information on all such recalls initiated 
and reported to FDA from January 1, 2005, through December 31, 2009. 
[Footnote 5] This information consisted of the most recent 5-year 
period of available data at the time we did our work. The source of 
this information was FDA's Recall Enterprise System (RES), the 
agency's central repository of recall information. FDA provided key 
information on each recall, including FDA's unique recall event 
number;[Footnote 6] the status of the recall at the time FDA provided 
us with this information (e.g., ongoing or terminated);[Footnote 7] 
the reason for the recall; the specific device being recalled; the 
recall classification level assigned based on FDA's assessment of 
risk; dates the recalls were initiated, classified, and terminated; 
and the medical specialty--area of use--for each device subject to 
recall (e.g., cardiovascular or orthopedic). We also obtained 
information identifying which of FDA's 19 district offices located 
throughout the United States were responsible for the day-to-day 
monitoring of the recalls. We then used this information to determine, 
among other things, the number of recalls initiated per year; the 
number of recalls by recall classification levels, the average length 
of time from initiation to termination for recalls that were 
terminated, and the number and percentage of recalls by medical 
specialty of the device being recalled. 

To assess the reliability of the information FDA provided, we reviewed 
FDA's user guide for RES and interviewed officials responsible for 
entering and reviewing the information in RES. Additionally, for a 
sample of recalls, we compared information from RES on the status and 
key dates to the source documents contained in FDA's recall files. We 
determined the data were sufficiently reliable for the purpose of our 
review. However, there are some limitations to our analyses. The data 
FDA provided may not include all recalls that firms actually initiated 
over this period. This is because FDA is dependent on firms self- 
reporting most recalls. In addition, FDA did not consistently enter 
certain data elements into RES over the 5-year period. For example, in 
certain cases the dates recalls were terminated were missing, and in 
many cases the root cause, or problem creating the need for the 
recall, was entered inconsistently. Therefore, we could not conduct 
certain analyses for all recalls initiated over the study period. 

In addition to identifying information about the numbers and 
characteristics of recalls, we identified the extent to which FDA uses 
recall information to aid its oversight of recalls. To accomplish 
this, we interviewed officials from FDA's Office of Regulatory Affairs 
(ORA), which develops FDA-wide policy on compliance and enforcement 
matters and also has primary responsibility for RES, the day-to-day 
monitoring of individual recalls, and conducting inspections of firms. 
We also interviewed officials from FDA's Center for Devices and 
Radiological Health (CDRH), which regulates medical devices marketed 
in the United States and is responsible for classifying recalls and 
assessing the adequacy of a firm's actions to correct problems leading 
to the most serious recalls. In addition, we interviewed 
representatives from two device manufacturer associations and several 
device manufacturers to obtain their views about the recall process. 

To identify the extent to which the medical device recall process 
ensures the effective implementation and termination of the highest- 
risk device recalls, we reviewed key documentation related to the 
recalling firms' management and FDA's oversight of a sample of the 
highest-risk recalls. We identified 131 recalls initiated over the 5- 
year period to which FDA assigned its highest-risk classification-- 
recalls for which FDA determined that there was a reasonable 
probability that the use of or exposure to the devices would cause 
serious adverse health consequence or death. Of these 131, we selected 
a sample of 53 recalls for in-depth review by identifying the four FDA 
district offices that had the largest number of available recall files 
to review. These 53 recalls represented all such recalls during the 5- 
year period that were managed by these four offices and for which 
there were files available for our review. For these 53 recalls, we 
obtained from FDA and reviewed the recall case files that were 
maintained by district offices. These files contained key documents 
such as information from the firms on the causes of the recalls, the 
firms' actions to prevent recurrence of similar problems, the recall 
notifications firms sent out to customers, FDA's correspondence with 
firms, and documentation from RES. As part of our review, we 
identified actions recalling firms took to adhere to FDA's guidance 
and regulatory requirements, and the extent to which the firms were 
able to recall all of the affected devices. We also identified whether 
FDA followed its own procedures for overseeing and terminating the 
recalls and whether FDA's recall process was in conformance with 
internal control standards for the federal government.[Footnote 8] For 
example, we reviewed the case files to determine whether or not FDA 
conducted recall audit checks in which FDA would contact a percentage 
of parties affected by the recall to determine whether they received 
the recall notice and followed the firms' instructions for removing or 
correcting the device. This included a review of over 2,000 audits 
check forms FDA completed for the 53 recalls we reviewed. Between 
December 2010 and February 2011 we discussed the recalls with 
officials from the responsible FDA district office, and determined the 
current status of the recalls. 

We conducted this performance audit from January 2010 to June 2011, in 
accordance with generally accepted government auditing standards. 
Those standards require that we plan and perform the audit to obtain 
sufficient, appropriate evidence to provide a reasonable basis for our 
findings and conclusions based on our audit objectives. We believe 
that the evidence obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives. 

Background: 

FDA is responsible for ensuring that medical products--including 
medical devices--sold in the United States provide reasonable 
assurance of safety and effectiveness and do not pose a threat to 
public health. FDA's oversight responsibilities for medical devices 
begin before a product is brought to market and continue after a 
product is available for sale. Its premarket responsibilities include 
reviewing thousand of submissions for new devices filed each year to 
decide whether they should be allowed to be marketed in the United 
States. Its postmarket responsibilities include monitoring the safety 
of thousands of medical devices already on the market and identifying, 
analyzing, and acting on potential risks the devices may pose to the 
public. This monitoring includes overseeing recalls of medical devices. 

FDA's Classification and Approval or Clearance of Medical Devices: 

FDA classifies each device type into one of three classes--class I, 
II, or III--based on the level of risk it poses and the controls 
necessary to provide reasonable assurance of its safety and 
effectiveness. According to FDA, the risk the type of device poses to 
the user is a major factor in the class it is assigned: class I 
includes devices with the lowest risk, and class III includes devices 
with the highest risk. Examples of types of devices in each class 
include the following: 

* class I: tongue depressors, elastic bandages, reading glasses, and 
forceps; 

* class II: electrocardiographs, powered bone drills, and mercury 
thermometers; and: 

* class III: pacemakers and replacement heart valves. 

In general, unless exempt under FDA regulations, medical devices are 
subject to one of two types of FDA premarket review before they may be 
legally marketed in the United States.[Footnote 9] These reviews are 
as follows. 

* Premarket approval (PMA): The manufacturer must submit evidence, 
typically including clinical data, providing reasonable assurance that 
the new device is safe and effective. The PMA process is the most 
stringent type of premarket review. A successful submission results in 
FDA's approval to market the device. 

* Premarket notification (510(k)): Premarket notification is commonly 
called "510(k)" in reference to section 510(k) of the Federal Food, 
Drug, and Cosmetic Act where the notification requirement is listed. 
Under this review, the manufacturer must demonstrate to FDA that the 
new device is substantially equivalent to a device already legally on 
the market.[Footnote 10] For most 510(k) submissions, clinical data 
are not required and substantial equivalence will normally be 
determined based on comparative descriptions of a device's intended 
use and technological characteristics, and may include performance 
data.[Footnote 11] A successful submission results in FDA's clearance 
to market the device. 

Most class I device types and some class II devices are exempt from 
FDA's premarket review. In general, those that are not exempt, but 
which are substantially equivalent to a legally marked class I or 
class II device, are subject to premarket review through the 510(k) 
process. Class III device types are generally required to obtain FDA 
approval through the more stringent PMA process.[Footnote 12] 

The Voluntary Medical Device Recall Process: 

FDA defines a recall as a firm's removal or correction of a marketed 
product that FDA (1) considers to be in violation of the laws it 
administers, and (2) against which the agency would initiate legal 
action.[Footnote 13] Nearly all medical device recalls are voluntarily 
initiated by a firm, usually the manufacturer of the device. The 
recall process generally consists of a series of steps that we have 
categorized into broad phases--initiating and classifying the recall, 
conducting and overseeing the recall, and completing and terminating 
the recall. While the recalling firm has primary responsibility for 
ensuring that the recalled devices are corrected or removed, FDA and 
other stakeholders each have responsibilities which they are supposed 
to undertake in order to effectively implement the various phases of a 
recall. FDA's role is generally to oversee a firm's management of 
recalls. It conducts its responsibilities as part of its postmarket 
surveillance. FDA staff from ORA--which is the lead office for all FDA 
field activities, including the agency's district offices--and CDRH 
are involved in overseeing recalls. Other stakeholders, including the 
firm's customers--such as distributors--and device users--such as 
hospitals or patients--are expected to correct or remove the recalled 
device according to the recalling firm's instructions.[Footnote 14] A 
given recall may require the cooperation of thousands of different 
stakeholders depending on how many entities received, purchased, or 
used the device. 

The following sections generally describe the voluntary recall process 
that FDA, as well as recalling firms, their customers, and device 
users, are expected to follow according to FDA's regulations, 
procedures, and guidance. 

Initiating and Classifying the Recall: 

During this phase of a device recall, a firm initiates a recall, while 
FDA classifies the recall based on health risks presented by use of 
the device. As part of this phase, a firm develops a strategy for 
implementing the recall, and FDA reviews and suggests changes to the 
strategy. In most cases, a firm arrives at the decision to initiate a 
recall after discovering a problem with a device, or a series of 
similar devices.[Footnote 15] The firm may then contact an FDA 
district office or immediately begin implementing a recall.[Footnote 
16] A firm may initiate a recall--that is notify stakeholders such as 
distributors and device users about the recall--prior to contacting 
the FDA district office. However, according to federal regulations, a 
firm must provide FDA with a report of correction or removal within 10 
working days of initiating a recall of a product that involves or may 
involve a risk to health.[Footnote 17] As part of its report, the firm 
is to provide FDA with key information such as the reason the device 
is being recalled, the brand name and model of the device, the lot or 
serial numbers of the device, the number of devices subject to 
correction or removal, and contact information for its customers and 
device users who received, used, or purchased the device. According to 
FDA's guidance, the recalling firm is also asked to develop a recall 
strategy that takes into account its assessment of the health hazard 
associated with the device. The strategy should contain details on the 
firm's plan for ensuring that its customers and device users correct 
or remove the device according to the firm's instructions, and the 
need for public warnings about the device. As part of its oversight, 
FDA will review the strategy, and may suggest that the firm make 
changes to its approach for conducting the recall. 

Once the district office is notified about the recall, it should 
create a record in RES, notify CDRH, and obtain and evaluate 
information CDRH needs to make its classification decision. The 
district office monitoring the recall will provide any information it 
receives from the firm, including the correction and removal report, 
to CDRH so it can begin the process of classifying the recall. For 
some recalls, the district office may need to conduct a recall 
inspection at the establishment where the device is manufactured in 
order to obtain additional information needed to classify the recall. 
According to FDA's procedures, when a recall appears to involve 
significant health risks, an inspection should be conducted to 
determine, among other things, the root causes of the problem and if 
the firm is implementing appropriate corrective action. The inspection 
may be performed by the FDA district office monitoring the recall or 
other district offices, such as those located near the firm's 
manufacturing establishment. 

To classify the recall, CDRH is to conduct its own health risk 
assessment of the device being recalled. Based on this assessment, 
CDRH classifies the recall to indicate the relative degree of health 
hazard presented by use of the device. According to CDRH's procedures, 
the classification decision should be completed within 31 calendar 
days from the time it received the information from the district 
office.[Footnote 18] Recalls are classified into one of three 
categories: 

* class I--reasonable probability that the use of, or exposure to, a 
device will cause serious adverse health consequences or death. These 
are the most serious recalls. 

* class II--use of or exposure to a device may cause temporary or 
medically reversible adverse health consequences, or the probability 
of serious adverse health consequences is remote. 

* class III--use of, or exposure to, a device is not likely to cause 
adverse health consequences.[Footnote 19] 

Table 1 compares FDA's classification of medical devices and recalls 
according to risk. It is important to note that FDA's device and 
recall classification schemes carry opposite designations. The 
potential degree of health risk associated with device classes is 
designated from class III (high) to class I (low), while the potential 
risk associated with recall classes is designated from class I (high) 
to class III (low). 

Table 1: Comparison of FDA Medical Device Classification and Recall 
Classification Levels, by Risk: 

Risk: High; 
Device classification: III; 
Recall classification: I. 

Risk: Moderate; 
Device classification: II; 
Recall classification: II. 

Risk: Low; 
Device classification: I; 
Recall classification: III. 

Source: GAO analysis of FDA classification information. 

[End of table] 

Once the recall is classified, FDA is to notify the firm, in writing, 
of the assigned recall classification. This classification letter 
should also include instructions about the extent to which the firm 
should conduct effectiveness checks--that is, contacting customers and 
device users to determine whether the recall notification was received 
and acted upon appropriately. In general, for class I recalls, FDA 
recommends that firms conduct effectiveness checks with 100 percent of 
customers and device users affected by the recall. For class II 
recalls, FDA recommends effectiveness checks with 10 percent of such 
customers and device users, and 2 percent for class III recalls. 

Conducting and Overseeing the Recall: 

During this phase, the firm and recall stakeholders are supposed to 
implement the recall as outlined in the approved recall strategy, and 
FDA is responsible for monitoring the progress made. Once a recall is 
under way, the firm is to conduct effectiveness checks to ensure that 
those stakeholders affected by the recall have received notification 
about the recall and have taken appropriate action, such as returning 
defective devices, or taking actions to correct the known defects. 
(See appendix I for information on tools to help customers and medical 
device users identify recalled devices and an FDA initiative intended 
to better track devices through the use of unique identifiers.) 
Additionally, at the request of the FDA district office responsible 
for monitoring the recall, the firm is expected to provide status 
reports on the progress of the recall. These reports should include 
information on how many customers and device users have received the 
recall notification and followed the firm's instructions, and how many 
still need to respond to the recall notice. The FDA district office 
reviews the reports, and, using RES, assigns the recall a status of 
ongoing if the reports indicate the recall is still under way. 

During the recall, FDA district offices independently assess the 
effectiveness of the recall by conducting audit checks.[Footnote 20] 
According to the agency's procedures, for each check, investigative 
staff from one or more of FDA's district offices will contact 
individual distributors or device users. These audit checks are 
generally conducted in person or by telephone, to confirm that the 
distributor or device user (1) received notification from the firm 
about the recall and (2) properly corrected or removed the recalled 
devices in accordance with the firm's recall strategy. The FDA 
district office responsible for monitoring the recall assigns the 
audit checks to one or more of the district offices, depending upon 
the location of the firm's customers and the device users. According 
to FDA procedures, the district office monitoring the recall should 
assign audit checks within 10 days of the recalling firm's initiation 
of the recall. The audit checks should be completed by FDA 
investigators, if possible, within 10 days of assignment. If an 
investigator determines that the firm and the distributor or device 
user followed the recall strategy, the investigator's audit check 
should conclude that the recall was effective. If not, the 
investigator's audit check should conclude that the recall was 
ineffective. The result of the audit check is documented on a 
standardized FDA form, and each form is provided to the district 
office that made the audit check assignment. 

Completing and Terminating the Recall: 

Once the firm believes it has completed the recall--i.e., done 
everything as outlined in the recall strategy--it needs to submit a 
final recall status report/recall termination request to the FDA 
district office monitoring the recall.[Footnote 21] Regardless of the 
class of the recall, if the district office agrees that the firm has 
completed the recall, it is to change the status of the recall in RES 
to completed. If it disagrees, it generally requests the firm to take 
additional actions, such as re-contacting customers and device users. 
The FDA district office bases its assessment of whether the recall has 
been effectively completed by reviewing the firm's status reports and 
results of the audit checks. In addition, according to FDA procedures, 
the final monitoring step the district office may take is to conduct a 
limited postrecall inspection to verify that the recall has been 
completed. During this inspection, investigators should witness 
destruction or reconditioning of the recalled product, if applicable. 

Once the district office considers a recall completed, FDA assesses 
whether it can terminate a recall. As part of its assessment, FDA may 
review a corrective and preventive action plan submitted by the 
recalling firm that describes the firm's actions to prevent a 
recurrence of the problem that led to the recall. Thus, this phase of 
the recall process is important because it provides FDA with the 
opportunity to determine whether the firm has taken sufficient 
corrective and preventive actions. The agency's procedures state that 
if a firm's corrective and preventive actions are adequate, FDA staff 
should terminate a recall within 3 months of completion. When 
terminating a class I recall, the district office sends a recall 
termination recommendation to CDRH. CDRH reviews the recalling firm's 
corrective and preventive action plan, and effectiveness and audit 
check results, and makes the decision on whether to terminate the 
recall. The district office does not need CDRH approval to terminate 
class II and III recalls. If corrective actions are determined 
sufficient, the recall status in RES is changed from completed to 
terminated. When FDA terminates a recall, the district office will 
close the recall file and notify the firm, in writing, that it can 
cease recall activity. Figure 1 displays the general process from 
initiating to terminating a recall. 

Figure 1: The Voluntary Medical Device Recall Process: 

[Refer to PDF for image: illustration] 

Initiation and classification: 

The firm: 
* identifies a need for and initiates a device recall,
* contacts an FDA district office, 
* develops a recall strategy. 

FDA: 
* the district office contacts CDRH and provides preliminary 
information, 
* CDRH classifies the recall and informs the recalling firm, 
* the district office and CDRH review and approve the firm’s recall 
strategy. 

Conducting and overseeing the recall: 

The firm: 
* provides customers and device users with instructions for implementing
the recall, 
* follows up with customers and device users and updates FDA with recall
progress. 

Customers and device users: 
* such as distributors, hospitals, retailers, and individual consumers
follow the recall instructions for correcting or removing the device. 

FDA: 
* the district office monitors the recall by reviewing the firm's 
status reports, 
* assesses recall effectiveness by performing audit checks. 

Completion and termination: 

The firm: 
* indicates that the recall is complete and requests that it be 
terminated. 

FDA: 
* determines whether a recall is complete and whether it should be 
terminated. 

Source: GAO analysis of FDA information. 

Note: According to FDA officials, the various phases of the recall 
process may not necessarily occur sequentially, and individual actions 
may be occurring simultaneously in different phases of the process for 
specific recalls. 

[End of figure] 

Firms Initiated Several Thousand Medical Device Recalls, but FDA Has 
Not Routinely Analyzed This Information to Aid Its Oversight of 
Recalls: 

From 2005 through 2009, firms initiated 3,510 medical device recalls. 
Most of these were for medical devices in five areas of use or medical 
specialty areas. On average, the recall process took just over 420 
days from initiation to termination, with class I recalls (the highest-
risk recalls) averaging nearly 520 days. FDA has not routinely 
analyzed information about recalls to aid its oversight of the recall 
process, and thus could not explain trends in recalls over this time 
period. 

Several Thousand Device Recalls Were Initiated; Most Were Class II and 
Involved Cardiovascular and Radiological Devices: 

Between January 1, 2005, and December 31, 2009, firms initiated 3,510 
device recalls, an average of just over 700 per year. The annual 
volume fluctuated over this period, and ranged from a low of 658 in 
2006 to a high of 796 in 2008. FDA classified the vast majority of all 
recalls--nearly 83 percent--as class II, meaning use of these devices 
may cause temporary adverse health consequences (moderate risk). FDA 
classified 14 percent as class III, meaning use of the device is not 
likely to cause any adverse consequences (lowest risk); and 4 percent 
were classified as class I (highest risk), because FDA determined that 
there was a reasonable probability that the use of or exposure to a 
violative product would cause serious adverse health consequences or 
death (see figure 2). The number of class I recalls initiated between 
2005 and 2009 ranged from 17 to 41. For example, in 2007, 25 class I 
recalls were initiated; in 2008, 17 were initiated; and in 2009, 41 
were initiated. In comparison, the number of class II recalls 
generally increased each year and consistently exceeded 500 annually. 

Figure 2: Number of Medical Device Recalls Initiated from 2005 through 
2009: 

[Refer to PDF for image: stacked vertical bar graph] 

Recall initiation year: 2005; 
Class I recall (highest risk): 30; 
Class II recall (moderate risk): 133; 
Class III recall (lowest risk): 503; 
Total recalls: 666. 

Recall initiation year: 2006; 
Class I recall (highest risk): 18; 
Class II recall (moderate risk): 110; 
Class III recall (lowest risk): 530; 
Total recalls: 658. 

Recall initiation year: 2007; 
Class I recall (highest risk): 25; 
Class II recall (moderate risk): 91; 
Class III recall (lowest risk): 612; 
Total recalls: 728. 

Recall initiation year: 2008; 
Class I recall (highest risk): 17; 
Class II recall (moderate risk): 89; 
Class III recall (lowest risk): 690; 
Total recalls: 796. 

Recall initiation year: 2009; 
Class I recall (highest risk): 41; 
Class II recall (moderate risk): 57; 
Class III recall (lowest risk): 564; 
Total recalls: 662. 

Source: GAO analysis of FDA data. 

[End of figure] 

Our analysis found that approximately 60 percent of recalls during 
this period were for devices from five areas of use or medical 
specialty areas--cardiovascular, radiological, orthopedic, general 
hospital and personal use, and diagnostic chemistry.[Footnote 22] 
According to FDA, these medical specialties are among those with the 
greatest number of devices on the market and four of the five 
specialties--cardiovascular, radiological, orthopedic, and general 
hospital--account for the greatest number of devices cleared or 
approved for marketing each year. The remaining recalls were for 
devices in 19 other areas (such as general and plastic surgery and 
neurological devices); no other specialty accounted for more than 8 
percent of recalls (see table 2). 

Table 2: Number and Percentage of Medical Device Recalls Initiated 
from 2005 through 2009, by Medical Specialty Area: 

Medical specialty area: Cardiovascular; 
Total number of recalls: 532; 
Percentage of all recalls: 15; 
Number of class I recalls[A]: 40; 
Percentage of class I recalls: 31. 

Medical specialty area: Radiological; 
Total number of recalls: 484; 
Percentage of all recalls: 14; 
Number of class I recalls[A]: 2; 
Percentage of class I recalls: 2. 

Medical specialty area: Orthopedic; 
Total number of recalls: 410; 
Percentage of all recalls: 12; 
Number of class I recalls[A]: 4; 
Percentage of class I recalls: 3. 

Medical specialty area: General hospital and personal use[B]; 
Total number of recalls: 388; 
Percentage of all recalls: 11; 
Number of class I recalls[A]: 31; 
Percentage of class I recalls: 24. 

Medical specialty area: Diagnostic chemistry[C]; 
Total number of recalls: 315; 
Percentage of all recalls: 9; 
Number of class I recalls[A]: 16; 
Percentage of class I recalls: 12. 

Medical specialty area: Other specialty areas; 
Total number of recalls: 1,381; 
Percentage of all recalls: 39; 
Number of class I recalls[A]: 38; 
Percentage of class I recalls: 29. 

Medical specialty area: Total; 
Total number of recalls: 3,510; 
Percentage of all recalls: 100; 
Number of class I recalls[A]: 131; 
Percentage of class I recalls: 100. 

Source: GAO analysis of FDA data. 

Note: Percentages do not add to 100 percent due to rounding. 

[A] Class I recalls involve a reasonable probability that the use of, 
or exposure to, a device will cause serious adverse health 
consequences or death. 

[B] General hospital and personal use devices include bandages, 
examination gowns, and stretchers. 

[C] Diagnostic chemistry devices include test systems, such as those 
for measuring blood glucose. 

[End of table] 

As table 2 shows, for class I recalls, the greatest numbers were for 
devices from the cardiovascular medical specialty. In addition, the 
table shows that devices from the general hospital and personal use 
and diagnostic chemistry medical specialties accounted for a 
substantial number of class I recalls. Among class I recalls, we found 
that the largest number for cardiovascular devices involved automatic 
external defibrillators. The largest number of recalls for general 
hospital and personal use devices involved infusion pumps, including 
implantable programmable pumps. 

RES also contains information on the root cause of recalls, that is, 
the problem creating a need for the recall. On average, in 2008 and 
2009 (the only years for which FDA tracked these data in RES) the 
greatest numbers of recalls were caused by problems with manufacturing 
processes. FDA refers to this root cause as process control-- 
developing, conducting, controlling, and monitoring production 
processes to ensure that a device conforms to its specifications. 
Other leading causes were device design and software design. The two 
most common causes of class I recalls were the same as for all 
classes--process control and device design--but the third cause was 
component design or selection (see table 3). In general, FDA officials 
indicated they do not believe that there is a relationship between 
root cause and recall class. However, FDA officials indicated that 
some root causes of recalls are more likely to affect certain types of 
devices. For example, they stated that the root cause "incorrect or 
missing expiration date" is typically related to devices that involve 
sterilization. 

Table 3: Number and Percentage of Medical Device Recalls Initiated 
During 2008 and 2009, by Root Cause: 

FDA assigned root cause[A]: Process control[C]; 
Total number of recalls: 233; 
Percentage of all recalls: 16; 
Number of class I recalls[B]: 17; 
Percentage of class I recalls: 29. 

FDA assigned root cause[A]: Device design; 
Total number of recalls: 201; 
Percentage of all recalls: 14; 
Number of class I recalls[B]: 12; 
Percentage of class I recalls: 21. 

FDA assigned root cause[A]: Software design; 
Total number of recalls: 181; 
Percentage of all recalls: 12; 
Number of class I recalls[B]: 1; 
Percentage of class I recalls: 2. 

FDA assigned root cause[A]: Component design/selection; 
Total number of recalls: 59; 
Percentage of all recalls: 4; 
Number of class I recalls[B]: 7; 
Percentage of class I recalls: 12. 

FDA assigned root cause[A]: Additional root causes[D]; 
Total number of recalls: 782; 
Percentage of all recalls: 53; 
Number of class I recalls[B]: 21; 
Percentage of class I recalls: 36. 

FDA assigned root cause[A]: Total; 
Total number of recalls: 1,456; 
Percentage of all recalls: 100; 
Number of class I recalls[B]: 58; 
Percentage of class I recalls: 100. 

Source: GAO analysis of FDA data. 

[A] The root cause is the problem creating a need for the recall. 

[B] Class I recalls involve a reasonable probability that the use of, 
or exposure to, a device will cause serious adverse health 
consequences or death. 

[C] Process control entails developing, conducting, controlling, and 
monitoring production processes to ensure that a device conforms to 
its specifications. 

[D] Other root causes include false or misleading labeling, mistaken 
use of materials or components, or employee error. 

[End of table] 

Among all classes of recalls, we found that a higher proportion of 
recalls were for devices which were cleared for market through the 
510(k) process as compared to other FDA review processes. This 
reflects the fact that the overwhelming majority of devices--99 
percent, according to FDA--enter the market through this review 
process. Our analysis of RES data for 2,773 recalls[Footnote 23] found 
that 87 percent of recalls involved a device cleared through the 
510(k) process, nearly 8 percent involved a device approved through 
the more stringent PMA or PMA supplement process, and nearly 6 percent 
involved devices that were cleared through the 510(k) process and 
approved through the PMA process, or that were exempt from FDA review. 
[Footnote 24] 

We found similar trends for 101 class I recalls.[Footnote 25] We found 
that 74 of the recalls (73 percent) were for devices cleared through 
the 510(k) process, 22 percent were for PMA-approved devices, and the 
remaining 5 percent involved devices that were cleared through the 
510(k) process and approved through the PMA process, or that were 
exempt from FDA review.[Footnote 26] Compared to all recall classes, a 
higher percentage of class I recalls involved devices cleared through 
the PMA process (22 percent compared with 8 percent for all classes of 
recalls combined), which likely reflects the high risk of these 
devices. 

Additionally, we found that 14 of those 74 class I recalls involving 
devices that were cleared through the 510(k) process were for devices 
that FDA designated as high-risk devices--class III devices. We 
further found that all 14 recalls involved cardiovascular devices, 
including 12 for automatic external defibrillators. 

FDA Has Terminated Most Medical Device Recalls, with the Recall 
Process Taking over 500 Days for the Highest-Risk Recalls: 

At the time of our review, the 3,510 medical device recalls initiated 
from 2005 through 2009 were in various stages of the recall process. 
Approximately 60 percent--2,050--of all recalls initiated in this 
period had been terminated by FDA as of April 16, 2010, the date we 
received data from FDA. Firms had completed another 5 percent and were 
awaiting FDA's review and decision on termination. The remaining 36 
percent were ongoing (see figure 3). 

Figure 3: Status of Medical Device Recalls Initiated from 2005 through 
2009: 

[Refer to PDF for image: pie-chart] 

Terminated[A]: 59% (2,050); 
Ongoing: 36% (1,268); 
Completed[B]: 5% (192). 

Source: GAO analysis of FDA data. 

Note: Status as of April 16, 2010. 

[A] Terminated--FDA headquarters determined that firms' corrective 
actions taken were sufficient to prevent a recurrence of the problems 
that led to the recall. 

[B] Completed--An FDA district office concluded that the firm had 
essentially fulfilled their responsibilities for correcting or 
removing the devices. 

[End of figure] 

We found that for recalls that had been terminated, the time between 
the firm's initiation and FDA's termination of a recall varied by 
class.[Footnote 27] On average, over 420 days passed between 
initiation of a recall and FDA's termination. Among all classes of 
recalls, class I recalls took the longest--on average 516 days, while 
on average class II recalls took 436 days and class III took 352 days. 
[Footnote 28] The amount of time needed to conduct and terminate 
recalls was split roughly evenly between the portions of the process 
that are primarily the recalling firms' responsibilities--conducting 
the recall itself--and the portions that are primarily FDA's 
responsibility--oversight of the recall (see figure 4). 

Figure 4: Average Time from Initiation to Termination for Medical 
Device Recalls Initiated from 2005 through 2009: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Average number of days between initiation and completion: 231; 
Average number of days between completion and termination: 192; 
Total: 423. 

Class I: 
Average number of days between initiation and completion: 266; 
Average number of days between completion and termination: 250; 
Total: 516. 

Source: GAO analysis of FDA data. 

Note: Termination dates were available for 2,028 of the 2,050 recalls 
that were terminated during this period. 

[End of figure] 

FDA frequently did not meet its 3 month time frame for terminating 
completed recalls. It did not meet this time frame for more than half 
of all recalls and over 70 percent of class I recalls (see figure 5). 
On average, FDA took 192 days to terminate a recall after it 
determined a recall was completed, more than twice the time specified 
in its procedures. For class I recalls, the average was 250 days. FDA 
could not specifically identify reasons that explained why it took 
this amount of time to make termination decisions. The agency did 
indicate that termination time frames are affected by both FDA's 
ability to address recalling firms' termination requests, as well as 
firms' ability to provide adequate information in support of the 
termination decision. This information may include a sufficient 
corrective and preventive action plan to prevent a recurrence of the 
problem which led to the recall. These data indicate that the 
timeliness of recall termination decisions appears to have 
deteriorated since our 1998 report.[Footnote 29] 

Figure 5: Percentage of Medical Device Recalls Terminated More Than 90 
Days after Completion Date, for Recalls Initiated in Calendar Years 
2005 through 2009: 

[Refer to PDF for image: vertical bar graph] 

Recall class: Class I; 
Percentage of recalls exceeding FDA’s 90 day termination standard: 71%. 

Recall class: Class II; 
Percentage of recalls exceeding FDA’s 90 day termination standard: 51%. 

Recall class: Class III; 
Percentage of recalls exceeding FDA’s 90 day termination standard: 47%. 

Recall class: All classes; 
Percentage of recalls exceeding FDA’s 90 day termination standard: 51%. 

Source: GAO analysis of FDA data. 

Note: Completion and termination dates were available for 1,992 of the 
2,050 recalls that were terminated. A recall is considered complete 
when FDA agrees that the recalling firm has sufficiently completed all 
recall-related activities. A recall is terminated when FDA determines 
that in addition to completing the recall, the recalling firm has 
taken adequate corrective measures to prevent a recurrence of issues 
which led to the recall. 

[End of figure] 

At the time of our review, 36 percent--1,268 recalls--were ongoing. Of 
these, most had been initiated in the past few years; however, some 
have been ongoing since 2005, the beginning of our review period. Of 
those recalls that were ongoing, most were initiated in 2008 and 2009; 
however, 456 (36 percent) had been ongoing for at least 2 years, 
including 86 that had been ongoing for nearly 5 years. 

FDA Has Not Routinely Analyzed Data to Identify Systemic Problems 
Underlying Device Recalls: 

Although RES contains numerous data elements that would allow for 
analyses of recall data, FDA is not effectively using these data to 
identify whether there are systemic problems underlying recalls. 
[Footnote 30] Instead of using RES to conduct systemic analyses of 
recalls, which would be consistent with one of the agency's strategic 
goals--improving the quality and safety of manufactured products in 
the supply chain--FDA has used RES primarily for processing and 
tracking the progress of individual recalls. 

Agency officials have not been using RES as a management tool to 
conduct broad surveillance of recalls and related issues. Neither the 
district offices we contacted nor CDRH officials prepared routine 
reports that would enable officials to identify areas of potential 
concern in the recall process, such as recalls that have been ongoing 
for an extended period, or whether specific manufacturing or design 
problems are causing increases in recalls or the types of devices 
being recalled. In fact, FDA officials appeared to be unaware of RES's 
capability to generate summary data. When we requested data from RES, 
FDA staff were unable to extract these data themselves, and initially 
indicated that it would be impossible to obtain data from RES. After 2 
months, FDA officials concluded that through a special arrangement 
with a contractor they could obtain the RES data and meet our request. 

After we completed our analysis of the RES data, we provided key 
summaries to FDA officials, and asked them to comment on trends that 
we observed. Officials indicated that they have not fully analyzed 
these data and could not explain trends without extensive research of 
individual case files. They indicated that at most, they could offer 
speculation about some of the trends we observed. For example, they 
could not explain why the majority of recalls are class II, why class 
I recalls more than doubled between 2008 and 2009, or why many recalls 
had been ongoing for 5 years. Officials also could not provide 
definitive answers when we asked them to comment on other related 
topics, such as: 

* common causes of recalls; 

* trends in the number of recalls over time; 

* variation in the numbers of recalls by recall classification levels; 

* types of devices and medical specialties of devices accounting for 
most recalls; 

* the length of time needed for firms to complete recalls; and: 

* the length of time needed for FDA to terminate recalls. 

Although FDA has not been routinely analyzing recall data to identify 
whether there are systemic problems affecting recalls, officials 
indicated they have used these data to help direct their inspection 
resources, and to support compliance and enforcement actions. First, 
FDA officials indicated they use recall information as one of many 
elements to assess the relative risks that device manufacturers 
present, and thus which firms the agency should inspect in a given 
year. For example, the officials said that recall data is one of 
several elements that feed into a predictive model that determines the 
likelihood that firms are out of compliance with applicable laws or 
regulations, and therefore in need of inspection. Second, they told us 
they have plans to use recall information as the basis for developing 
a directed inspection plan. As part of this project, officials would 
use recall information to identify those firms that generate a large 
number of recalls, and target them for inspection. Officials indicated 
that these inspections would focus on specific areas--such as a 
particular manufacturing process. This effort is still in the planning 
phase, and officials have not yet established criteria, such as what 
constitutes a large number of recalls, for determining which firms to 
select. The officials also indicated that progress may be slow because 
they do not have sufficient resources available to devote to this 
effort.[Footnote 31] 

Although FDA has not regularly been using data to identify systemic 
problems, we found one example of FDA using recall data to detect and 
address safety issues with a particular type of device. In December 
2010, FDA held a conference on a variety of issues related to 
automatic external defibrillators, including the safety of these 
devices. During this conference it presented historical recall data to 
help demonstrate the need for a specific focus on safety improvements 
for this type of device.[Footnote 32] 

Gaps in the Medical Device Recall Process Limit the Effective 
Implementation and Termination of the Highest-Risk Recalls: 

Gaps in the medical device recall process limit firms' and FDA's 
ability to ensure that the highest-risk recalls are implemented 
effectively and terminated in a timely manner. We found that both FDA 
and recalling firms generally upheld their respective responsibilities 
in the course of initiating and classifying recalls. However, FDA did 
not always follow its own procedures and some procedures are unclear. 
FDA did not consistently inspect the manufacturing establishments of 
recalling firms as outlined in the agency's procedures. FDA has also 
not established criteria, such as thresholds, based on the nature of 
devices, for assessing whether firms effectively completed recalls by 
correcting or removing a sufficient number of recalled devices. 
Further, we found that firms face challenges, such as locating 
specific devices or users of devices, and often could not correct or 
remove all devices. We also found that audit checks, a key mechanism 
for FDA's oversight of firms' conduct of recalls, are limited in 
scope. In addition, because of a lack of clarity in FDA's audit check 
procedures, they have been implemented inconsistently by FDA's 
district offices. Finally, FDA frequently failed to make recall 
termination decisions in a timely manner, and kept no documentation to 
justify its termination decisions. 

Firms and FDA Have Generally Implemented Recalls Appropriately in the 
Initiation and Classification Phases, but FDA Has Failed to 
Consistently Conduct Recall-Related Inspections: 

In our review of a sample of the highest-risk device recalls initiated 
from January 1, 2005, through December 31, 2009, we found that once 
firms initiated recalls, they generally provided FDA with a correction 
or removal report in a timely manner--within FDA's 10-day time frame. 
In 51 of the 53 recalls (96 percent), firms submitted a correction or 
removal report to FDA. For 43 of these 51 recalls, firms submitted the 
report within 10 working days of initiating the recall. For 6 of the 
remaining 8 recalls, the correction and removal report was submitted 
within 21 business days; the reports for the other 2 recalls were 
submitted 62 business days and 227 business days after the recall was 
initiated, respectively. Table 4 shows the proportion of recalls, by 
district, where firms submitted these reports and whether they were 
submitted within 10 working days. 

Table 4: Number, Percentage, and Timeliness of Correction or Removal 
Reports Submitted for a Sample of Class I Medical Device Recalls 
Initiated from 2005 through 2009, by FDA District: 

District office: Detroit; 
Number of recalls: 9; 
Number of correction or removal reports submitted: 9; 
Percentage of correction or removal reports submitted: 100%; 
Number of correction or removal reports submitted within 10 working 
days: 8; 
Percentage of correction or removal reports submitted within 10 
working days: 89%. 

District office: Los Angeles; 
Number of recalls: 15; 
Number of correction or removal reports submitted: 14; 
Percentage of correction or removal reports submitted: 93%; 
Number of correction or removal reports submitted within 10 working 
days: 9; 
Percentage of correction or removal reports submitted within 10 
working days: 64%. 

District office: Minneapolis; 
Number of recalls: 16; 
Number of correction or removal reports submitted: 16; 
Percentage of correction or removal reports submitted: 100%; 
Number of correction or removal reports submitted within 10 working 
days: 15; 
Percentage of correction or removal reports submitted within 10 
working days: 94%. 

District office: New England; 
Number of recalls: 13; 
Number of correction or removal reports submitted: 12; 
Percentage of correction or removal reports submitted: 92%; 
Number of correction or removal reports submitted within 10 working 
days: 11; 
Percentage of correction or removal reports submitted within 10 
working days: 92%. 

District office: Total; 
Number of recalls: 53; 
Number of correction or removal reports submitted: 51; 
Percentage of correction or removal reports submitted: 96%; 
Number of correction or removal reports submitted within 10 working 
days: 43; 
Percentage of correction or removal reports submitted within 10 
working days: 84%. 

Source: GAO analysis of FDA medical device recall files. 

[End of table] 

Although our analysis indicates that firms generally provided these 
reports after initiating the recalls, FDA officials cautioned that 
this does not mean firms fully complied with the regulatory reporting 
requirements. They indicated that in some cases, firms' initial 
correction or removal reports lack some of the information needed and 
extra time was required for firms to provide additional information. 
To help address this, officials indicated that in November 2010 they 
began a recall process improvement project. As part of this 
initiative, CDRH plans to develop Web-based training modules for 
industry clarifying the information that needs to be provided when 
reporting corrections and removals to FDA.[Footnote 33] 

FDA infrequently--in less than one-half of the recalls--conducted an 
establishment inspection upon learning of a recall. According to FDA's 
procedures, upon learning of a potential class I recall, district 
offices should conduct establishment inspections to obtain further 
information about the recall. We found that FDA conducted such recall- 
related inspections for 20 of the 53 class I recalls we reviewed. 
[Footnote 34] The frequency of inspections varied across the four 
district offices monitoring the recalls. Three of these offices 
(Detroit, Los Angeles, and New England) conducted recall-related 
establishment inspections upon the initiation of a recall between 25 
percent and 38 percent of the time, while the Minneapolis district 
office conducted them in 62 percent of recalls (see figure 6). Based 
on interviews with FDA officials in four district offices, we found 
that decisions to conduct such inspections, given their overall 
inspection workload, are a matter of resources and timing. Some 
district officials also said that the decision to conduct a recall-
related inspection is based on the firm's recall history and indicated 
that FDA may be less likely to inspect a firm with a history of 
completing recalls successfully. Finally, some of these officials FDA 
said that this is because firms that have successfully completed 
recalls generally provide the necessary information, such as 
determinations of the root cause of the recall, as part of their 
correction or removal reports. 

Figure 6: Percentage of Our Sample of Class I Medical Device Recalls 
Initiated from 2005 through 2009 for Which Recall-Related Inspections 
Were Conducted, by FDA District Office: 

[Refer to PDF for image: vertical bar graph] 

FDA district office: Detroit; 
Percentage of recalls with inspections: 25%. 

FDA district office: Los Angeles; 
Percentage of recalls with inspections: 33%. 

FDA district office: Minneapolis; 
Percentage of recalls with inspections: 62%. 

FDA district office: New England; 
Percentage of recalls with inspections: 38%. 

Source: GAO analysis of FDA medical device recall files. 

Note: We selected a sample of 53 recalls for in-depth review by 
identifying the four FDA district offices that had the largest number 
of available recall files to review. This figure includes 49 of the 53 
recalls in our sample of class I recalls. FDA identified the problem 
that led to the remaining 4 recalls--1 from the Detroit and 3 from the 
Minneapolis district offices--through postmarket establishment 
inspections, unrelated to recalls. Therefore, FDA did not conduct 
additional inspections upon the initiation of these recalls. 

[End of figure] 

FDA generally followed its procedures by classifying each of the 53 
recalls in our sample and providing written notification to the 
recalling firms. However, for 28 of the 53 recalls, FDA did not make 
its classification determination within 31days as outlined in its 
procedures.[Footnote 35] The amount of time from recall initiation to 
classification varied, ranging from a few days to several months, with 
an average of 47 days. 

Representatives from two device manufacturer associations and several 
device manufacturers expressed concern about the length of time it can 
take FDA to classify recalls. For a class I recall, firms must make 
greater efforts to identify and contact customers than for class II 
recalls. Thus, delays in FDA's classification can affect firms' 
decisions. For example, officials indicated that if they send out a 
recall notice that they believe will be a class II, and after a 
significant amount of time FDA informs them it is a class I recall, 
the firm will have to revise the notice to indicate that the risks 
posed by the recall were more severe than they initially anticipated. 
The firm will also have to identify additional customers and device 
users to contact, in order to meet FDA's recommendation that they 
conduct 100 percent effectiveness checks for class I recalls. Firm 
officials said that they will then send out the revised notice, which 
can create confusion about whether this is a new recall or whether it 
is an update with new instructions for the already ongoing recall. 

Firms Faced Challenges in Correcting or Removing All Recalled Devices: 

Our review of firms' action in conducting recalls found that the 
status of recalls varied and that firms face challenges in correcting 
or removing all recalled products. Of the 53 recalls we reviewed, we 
found 13 were ongoing, 10 were completed--meaning that an FDA district 
office concluded that the firm had essentially fulfilled their 
responsibilities for correcting or removing the devices--and 30 were 
terminated--meaning FDA headquarters determined that firms' corrective 
actions taken were sufficient to prevent a recurrence of the problems 
that let to the recall (see figure 7).[Footnote 36] 

Figure 7: Status of Our Sample of 53 Class I Medical Device Recalls 
Initiated from 2005 through 2009: 

[Refer to PDF for image: pie-chart] 

Terminated[A]: 57% (30); 
Ongoing: 25% (13); 
Completed[B]: 19% (10). 

Source: GAO analysis of FDA medical device recall files. 

Note: We obtained the current status of these recalls through our 
reviews of FDA's recall files and discussions with FDA district office 
officials between December 2010 and February 2011. Percentages may not 
sum to 100 because of rounding. 

[A] Terminated--FDA headquarters determined that firms' corrective 
actions taken were sufficient to prevent a recurrence of the problems 
that let to the recall. 

[B] Completed--An FDA district office concluded that the firm had 
essentially fulfilled their responsibilities for correcting or 
removing the devices. 

[End of figure] 

Of the 40 recalls in our sample that were either completed or 
terminated--meaning that FDA concluded that the firm had taken 
sufficient effort to correct or remove recalled devices--we found that 
for 19 (48 percent) of these recalls, firms were able to correct or 
remove all products. In the other 21 recalls (53 percent) firms were 
unable to correct or remove all products. These recalls ranged widely, 
in both volume of devices subject to recall and the types of devices 
being recalled. Some recalls involved hundreds of thousands of 
disposable products, while others involved a small number of life- 
sustaining implantable devices. Although recalling firms took steps to 
notify customers and device users, they were often unable to correct 
or remove all devices. This was because firms could not locate some of 
the customers or device users, or these customers or device users 
could not locate the device subject to recall. In other cases this was 
because the devices had been disposed of (such as defective syringes), 
or were sold at retail outlets (such as glucose test strips) to 
individuals who may not have known about the recall. For example, in a 
recall of tracheal tubes included in certain pediatric medical kits, 
1,400 tubes had been distributed, but only 200 were returned to the 
recalling firm. The firm said that the rest had likely been used. 
Finally, users occasionally were unwilling to return a device. For 
example, one recall involved a magnetic device designed to treat a 
variety of medical problems such as lower back pain, fibromyalgia, and 
arthritis. This device was never cleared or approved by FDA, and 
despite FDA warnings about the device, users who had purchased units 
refused to return them. Details concerning the 21 recalls for which 
firms were not able to correct or remove all devices are presented in 
appendix II. 

FDA's Oversight of Recalls Is Inconsistent and Narrow in Scope and Its 
Procedures Are Unclear: 

Our review of FDA's actions for conducting and overseeing recalls 
revealed that FDA generally conducted audit checks for the class I 
recalls we reviewed, but we found unclear procedures led to numerous 
inconsistencies in how different investigators conducted these checks 
and made their determinations about the effectiveness of recalls. FDA 
conducted audit checks for 45 of the 53 recalls (85 percent) we 
reviewed.[Footnote 37] Our analysis of 2,196 audit check forms 
associated with these recalls found that audit checks completed for 
nearly 90 percent of the recalls contained a variety of 
inconsistencies in how the audit checks were implemented and 
documented.[Footnote 38] For each of these recalls we found 
inconsistencies in how different investigators determined whether a 
recall was effective or ineffective when conducting their audit checks 
of recalls. We also identified inconsistencies in the level of detail 
provided in the audit check report, and the level of effort undertaken 
by different investigators. Specifically, we found the following. 

* Some investigators' audit checks concluded that recalls were 
effective, despite noting problems (such as device users not following 
the firm's instructions), while other investigators concluded that 
similar instances were ineffective. For example, in 2008 a firm 
initiated a recall of an implantable pump because of problems in the 
connection between a catheter and the pump, which could result in 
improper amounts of medication being delivered to a patient. The 
firm's recall notification alerted physicians to this problem, and 
provided instructions for monitoring patients who already had the 
implanted pump and for revising future implant procedures. As part of 
the audit check program for this recall, FDA's investigators contacted 
a sample of physicians to determine whether they received the 
notification and followed the instructions. Our review found that out 
of 68 audit checks, there were 14 instances where the investigators 
noted that physicians either did not receive the recall notification, 
or did not remember receiving it, and thus could not have followed the 
recall notice instructions. In 8 of these 14 instances, investigators 
concluded that the recall was ineffective, noting that the physicians 
did not implement the recall instructions. In contrast, in the other 6 
instances they concluded the recall was effective, even though 
physicians could not have followed the recall instructions. In some 
cases this was because the firm provided evidence that they had 
notified the physician, and in others the investigator noted that the 
physician did not have any pumps on hand. 

* Some investigators determined that device users were not notified of 
the recall by the recalling firm, but instead learned of the recall 
through other means. In some of these instances, investigators' audit 
checks concluded that recalls were effective, while in other similar 
cases investigators concluded the checks were ineffective. 

* Some investigators wrote detailed comments on the audit check form 
as to why the investigator determined the recall was effective or 
ineffective, while others did not. Without comments, it may be 
difficult for FDA supervisors and district recall coordinators to 
verify whether an investigator correctly determined whether the recall 
was effective or ineffective. 

* Some investigators noted actions they took when they discovered 
problems with recalls, such as providing the device users with a copy 
of the recall notice or instructing them on actions to take in order 
to implement a recall. In contrast, other investigators did not 
indicate whether they made any attempt to help facilitate the recall. 
For example, in 2009 a firm initiated a recall of an automated 
external defibrillator because of reports that some of these devices 
failed to discharge sufficient energy due to problems with batteries. 
The firm issued a notice that instructed users to replace batteries 
and update software for the devices. As part of the audit checks for 
this recall, FDA investigators contacted a sample of users of the 
device, to check whether they received the recall notification and 
followed the firm's recall instructions. Our review found that out of 
67 audit checks, there were 35 instances where investigators noted 
problems with the recall--generally that the user did not receive the 
notice or failed to follow recall instructions. In 29 of these cases, 
the FDA investigator noted taking actions, including providing the 
recall notice or instructing the user to contact the recalling firm so 
they could obtain software needed to perform the needed actions. 
However, in 6 cases we found no indication that the FDA investigator 
took actions to ensure the recall was carried out effectively. 

FDA officials at both ORA and the district offices we contacted 
acknowledged that there are no detailed instructions or requirements 
for conducting audit checks, and that there can be inconsistencies in 
the process. Officials told us that when determining whether or not a 
check is effective, investigators should be assessing whether the 
recalling firm provided the notice and instructions to the customers 
or device users, and whether the customers or users followed 
instructions. They acknowledged, however, that some investigators may 
approach these checks differently, and that this may be an area where 
clarification of the agency's procedures is needed. During our 
interviews with officials from the Detroit, Los Angeles, Minneapolis, 
and New England district offices, some officials said that audit 
checks are typically conducted by new investigators, and that 
investigators receive classroom and on-the-job training on how to 
conduct such checks. Some district officials also noted that audit 
checks are reviewed by a supervisor as well as the recall coordinator 
in the district office that is monitoring the recall, and this serves 
as a quality control function to ensure consistency. Also, officials 
from FDA headquarters and some district offices stated that they have 
attempted to institute measures to improve the audit check process. 
Specifically, they noted that they recently updated the audit check 
form to more precisely reflect what makes a recall ineffective. Also, 
ORA officials indicated that they plan to automate the audit check 
forms, which will make the forms accessible to officials in FDA's 
headquarters. FDA officials said that they are considering 
applications for analyzing the automated data, but have not completed 
any specific plans. 

In addition to the inconsistencies, we found other gaps in FDA's 
oversight related to the audit checks. First, FDA's audit checks were 
often narrow in scope, in that FDA instructs investigators to contact 
only a small number of customers or device users--between 2 percent 
and 10 percent of those affected by the recall. Therefore, if there 
are thousands of customers or device users, the audit checks provide 
FDA with a means to contact a relatively small number of them. For the 
45 recalls for which FDA completed checks, we found the number of 
audit checks conducted varied widely, from 2 to 271, with an average 
of 51 audit checks per recall.[Footnote 39] Second, FDA investigators 
did not always conduct the assigned number of audit checks. We 
compared the number of audit checks that should have been conducted 
based on the audit check assignments to the numbers of checks actually 
completed. We found that for 17 of the 45 recalls (38 percent) fewer 
than the assigned number were conducted. Third, even though most 
checks were done in person, consistent with FDA's procedures, over 22 
percent of the checks were done by telephone. In these cases, the 
audit check relied extensively on anecdotal information provided by 
the customer or device user. According to FDA, the number of checks it 
can perform is limited by available resources. Based on our review of 
files, we found that if patients or consumers are involved (e.g., if 
FDA needed to contact someone with an implantable device), these were 
often done by telephone. We also found checks that were done by 
telephone for other device users including hospitals, retailers, and 
doctors' offices. 

Gaps in FDA's Completion and Termination Process Increase the Risk 
That Unsafe Medical Devices May Continue to Be Used: 

We found FDA lacks specific criteria for making decisions about 
whether recalling firms have adequately completed their recalls--a key 
oversight activity of the recall process. FDA officials indicated they 
consider a recall complete when a firm has completed actions outlined 
in its recall strategy. In particular, they evaluate whether firms 
completed their assigned level of effectiveness checks, and have 
corrected or removed recalled devices in "an acceptable manner." 
However, our review of FDA's recall procedures found--and FDA 
officials confirmed--that the procedures do not contain any specific 
criteria or general guidelines governing the extent to which firms 
should be correcting or removing various types of devices before a 
recall should be considered completed. For example, FDA does not have 
a benchmark recovery rate or threshold to assess whether firms 
effectively completed recalls, although the recovery rates of devices 
could be expected to vary, depending on whether a recalled device was 
a large piece of hospital equipment or a disposable device, such as a 
syringe. Representatives from medical device firms stated that there 
are no criteria or guidance from FDA on the percentage of recalled 
products that must be corrected or removed. Further, these firm 
representatives said that FDA is generally satisfied with three 
attempts at communicating with customers and device users affected by 
the recall. 

In addition, for a majority of the class I recalls we reviewed, FDA's 
actions to ensure that recalls were complete were inconsistent with 
its procedures for overseeing recalls. According to FDA's procedures, 
districts should conduct a limited postrecall inspection to verify 
that the recall is complete, and to witness destruction of defective 
products, if applicable. In 21 of the 40 completed and terminated 
recalls (53 percent) we found no documented evidence that FDA took 
actions besides audit checks to verify that the recall was complete. 
In the other 48 percent of recalls, FDA made an assessment via 
inspection, witnessing destruction of devices, or verifying that 
software corrections were completed. 

Another gap we found in the recall process is that FDA does not 
maintain sufficient documentation to justify its termination 
decisions. Although FDA may request that firms submit corrective and 
preventative action plans for review and approval before a recall can 
be terminated, we found little documentation on how FDA assessed 
whether such plans were sufficient when it terminated recalls. When we 
asked to review documentation justifying the decisions for the 
terminated recalls in our sample, FDA officials indicated that they do 
not maintain extensive documentation justifying the basis for their 
termination decisions. They told us that creating documentation to 
support concurrence with the termination recommendation is not part of 
past or current termination procedures. This approach is inconsistent 
with internal control standards for the federal government, which 
indicate "that all transactions and other significant events need to 
be clearly documented" and stress the importance of "the creation and 
maintenance of related records which provide evidence of execution of 
these activities as well as appropriate documentation."[Footnote 40] 
Without such documentation, we were unable to assess the extent to 
which FDA's termination process appropriately evaluated recalling 
firms' corrective actions. 

Also, we found that FDA termination decisions were frequently not made 
in a timely manner--within 3 months of the completion of the recall-- 
increasing the risk that unsafe or defective devices remained 
available for use. Of the 53 files in our sample, 30 were terminated--
meaning FDA headquarters determined that firms developed sufficient 
corrective actions to prevent a recurrence of problems which led to 
the recalls. For 73 percent of the terminated recalls, FDA did not 
make its termination decision within 3 months of the recall's 
completion, as indicated by FDA procedures. Overall, termination 
decisions took between 10 and 800 business days from completion to 
termination, with an average of 187 business days. Failure to make 
termination decisions in a timely manner increases the risk that 
patients and healthcare providers may continue to use unsafe or 
defective devices. For example, one firm requested termination from 
FDA for its recall of a portable external defibrillator in February 
2006. However, FDA did not begin its termination assessment until May 
2010. In this case, officials indicated that, due to staff turnover in 
the district office, they were unaware that this recall was still 
ongoing until a new recall coordinator searched for ongoing recalls. 
In 2010, following an FDA inquiry, the firm stated that it had not 
received confirmation of a required upgrade from 91 end users and an 
additional 13 devices could not be located. Because FDA did not follow 
up on this recall until 2010, 4 years had elapsed before the agency 
became aware that the recalling firm had not corrected or removed a 
substantial number of devices subject to the recall. According to FDA 
officials, their ability to terminate recalls in a timely manner is 
affected by resources, and termination decisions are a lower priority 
than other issues because the recalling firm has completed its actions. 

We found at least one instance where FDA's failure to make a timely 
termination assessment allowed for a potentially unsafe product to be 
reintroduced into the market and used for surgical procedures. In this 
case, based on adverse event reports that screws in its spinal 
fixation system were becoming loose postoperatively, the firm decided 
to recall the device in December 2005. The firm implemented its recall 
and removed all devices. The firm indicated that it developed a 
corrective action for the screw problem, and relaunched the device in 
April 2006. It then requested termination from FDA in May 2006. FDA 
followed up on this request by leaving three voice mail messages with 
the firm, and received no response. The agency sent out a request for 
information a year later, in May 2007. In June 2007, the company again 
indicated that the recall was complete, and requested termination. In 
September 2007, FDA conducted an inspection of the company's 
manufacturing facility, and found that while the recall was complete, 
the corrective action was not adequate. Over the course of the next 2 
years, the firm worked with FDA to get revisions to the device 
approved, but eventually agreed to a second recall for the revised 
device. This recall was initiated in May 2009. We identified five 
reports of adverse events related to continuing problems with the 
implanted device that were filed with FDA subsequent to the firm's 
relaunch of the device in April 2006. These reports were filed from 
December 2006 through March 2007, and revealed that in all cases, 
patients required surgical intervention to correct or remove the 
device. 

Conclusions: 

The medical device recall process is complex, requiring the 
coordination and timely action of potentially thousands of parties. It 
is an important tool used by firms and FDA to protect the public and 
mitigate health risks from unsafe or ineffective devices. While the 
recall process may not eliminate 100 percent of health risks 
associated with recalled devices, careful implementation and 
evaluation are critical to minimizing health risks. 

FDA has a key role in identifying and minimizing the public health 
risks presented by defective or unsafe devices. In this regard, FDA 
has opportunities to close some of the gaps that currently exist in 
the medical device recall process, and enhance its oversight of device 
recalls. As currently structured, FDA's approach to oversight of 
medical device recalls is reactive--responding to individual recalls 
as they occur. Rather than pursuing a strictly case-by-case approach 
to overseeing recalls, FDA could take a more proactive approach to its 
oversight. The agency has a plethora of data available on thousands of 
recalls, but at present, is not effectively reviewing and analyzing 
these data in a systematic manner. More routine analyses of these data 
could help FDA identify trends in the numbers and types of devices 
being recalled, as well as the underlying causes of device recalls. 
Such information would provide FDA with a better understanding of the 
risks presented by defective or unsafe devices, which could lead the 
agency to proactively identify strategies and measures needed to 
address systemic problems with the design or manufacture of individual 
devices or entire categories of devices. Armed with the results of 
these types of analyses, FDA could then be in a position to help 
mitigate safety risks before they occur, and thus minimize the need 
for recalls. This is particularly important for the devices involved 
in the highest-risk recalls, which place the public at risk of serious 
health consequences, including death. 

Furthermore, while the agency has devoted substantial resources to 
monitoring individual recalls, opportunities for enhancing its 
oversight of specific recalls also exist. A key FDA mechanism for 
overseeing individual recalls--audit checks of a small portion of 
customers and device users involved in the recall--are often 
implemented inconsistently. This is due to unclear procedures that 
investigators are using for implementing and documenting audit checks 
and making their final assessments. As a result, investigators can 
make inconsistent determinations about whether firms, customers, and 
device users have effectively conducted a recall. Additionally, FDA 
lacks clear criteria for determining whether firms have successfully 
completed recalls, and has failed to maintain important documentation 
justifying its decisions to terminate the highest-risk recalls. This 
impedes independent assessments of FDA's decision making and leaves 
the agency vulnerable to questions about the basis it used to 
determine that recalling firms fulfilled all their responsibilities 
when conducting recalls. By addressing these weaknesses, FDA could 
reduce the risk that defective or unsafe medical devices remain on the 
market, potentially endangering public health. 

Recommendations for Executive Action: 

To enhance FDA's oversight of medical device recalls, and in 
particular, those medical device recalls that pose the highest risk, 
we recommend that the Commissioner of FDA take the following four 
actions: 

* Create a program to routinely and systematically assess medical 
device recall information, and use this information to proactively 
identify strategies for mitigating health risks presented by defective 
or unsafe devices. This assessment should be designed, at a minimum, 
to identify trends in the numbers and types of recalls, devices most 
frequently being recalled, and underlying causes of recalls. 

* Clarify procedures for conducting medical device recall audit checks 
to improve the ability of investigators to perform these checks in a 
consistent manner. 

* Develop explicit criteria for assessing whether recalling firms have 
performed an effective correction or removal action. 

* Document the agency's basis for terminating individual recalls. 

Agency Comments: 

We provided a draft of this report to HHS for review. HHS's written 
comments are reprinted in appendix III. HHS agreed with our 
conclusions and recommendations and stated that the agency is 
committed to exploring each of our recommendations fully. HHS reported 
that FDA plans to convene a working group to both evaluate 
improvements to the recall process and to develop strategies to 
implement our recommendations. According to HHS, FDA recognizes that 
standardized guidance will strengthen the management of the recall 
process. In addition, HHS elaborated on some of FDA's efforts to 
enhance its oversight by, for example, developing more routine 
analysis and reporting of recall data. HHS also provided technical 
comments, which we incorporated as appropriate. We are greatly 
encouraged by the agency's response, and believe its expeditious 
implementation of the recommendations will serve to enhance the safety 
of medical devices used by millions of Americans each day. 

As agreed with your offices, unless you publicly announce the contents 
of this report earlier, we plan no further distribution until 30 days 
from the report date. At that time, we will send copies to the 
Commissioner of FDA and appropriate congressional committees. The 
report also will be available at no charge on GAO's Web site at 
[hyperlink, http://www.gao.gov]. 

If you or your staffs have any questions about this report, please 
contact me at (202) 512-7114 or crossem@gao.gov. Contact points for 
our Offices of Congressional Relations and Public Affairs may be found 
on the last page of this report. GAO staff who made major 
contributions to this report are listed in appendix IV. 

Signed by: 

Marcia Crosse: 
Director, Health Care: 

[End of section] 

Appendix I: Tools to Help Identify Recalled Devices and Planned 
Initiatives for Assisting with Device Recalls: 

Several key stakeholders involved in the medical device recall 
process, including recalling firms, device distributors, and device 
users--such as hospitals--share responsibilities for effectively 
implementing recalls. To implement an effective recall, stakeholders 
need mechanisms to ensure timely and open communication about the 
recalls, and a means of locating devices subject to recall. This 
appendix describes the recall notification and tracking systems 
available to help manage device recalls. It also provides information 
on the status of the Food and Drug Administration's (FDA) unique 
device identification (UDI) initiative--which is intended to enable 
the identification of a device throughout distribution and use. 

To obtain this information, we interviewed officials from FDA and key 
stakeholders, including representatives of firms providing 
subscription-based recall alert information, manufacturers, 
distributors, group purchasing organizations, hospital systems, and 
patient safety groups. Through these interviews we obtained 
information on what these stakeholders considered to be the key 
challenges they face in implementing device recalls. We also obtained 
information about mechanisms which, in particular, hospital systems 
use to help identify devices subject to recall. Further, we reviewed 
FDA's progress in implementing its UDI initiative. To accomplish this, 
we examined published studies on this initiative, reviewed stakeholder 
comments submitted for FDA's public meetings on the UDI, and 
interviewed FDA officials responsible for managing the UDI program. 
[Footnote 41] 

Many Stakeholders Use Electronic Recall Notification and Tracking 
Systems to Help Manage Device Recalls: 

We learned through stakeholder interviews that instead of relying 
solely on notifications from medical device manufacturers, health care 
providers and other stakeholders have come to rely on other sources of 
information to stay abreast of potential recalls. Furthermore, these 
electronic communication technologies are evolving and, over time, 
stakeholders have begun recognizing that they can play a role in 
helping to effectively implement recalls. 

Our interviews with stakeholders revealed that a number of different 
privately developed, subscription-based electronic notification and 
tracking systems are available to help identify and process recalls. 
Stakeholders indicated that these systems are primarily used by 
hospitals, but that the systems are available to others involved in 
recalls as well. Available services identify recalls from a number of 
sources, including device manufacturers and FDA's Web site. The 
services compile lists of recalls and send electronic messages about 
recalls to paid subscribers. These services vary in sophistication and 
price. One service we learned about was limited to periodic electronic 
notification of all recalls at a cost of about $500 per year. Others 
include software to help individual hospitals specifically delegate 
responsibility within their hospital system to specific officials who 
will manage certain aspects of the recalls--such as removing recalled 
products from inventory--and for tracking the progress of the recalls. 
These systems can cost several thousand dollars per year. Owners of 
these systems that we spoke to indicated that hundreds of hospitals 
subscribe to their systems (see table 5). 

Table 5: Examples of Available Recall Alert Services: 

Type of system: E-mail alerts; 
Service provided: Sends e-mail alerts to hospitals every 7-10 days or 
immediately for urgent recalls; 
Primary customers: Hospitals; 
Source of information in alerts: FDA, manufacturers, and subscribers. 

Type of system: Weekly newsletter, Web-based alerts and tracking; 
Service provided: An alert service and optional Web-based recall and 
safety alert tracking system that helps clients to acquire, 
distribute, and manage recalls; 
Primary customers: Primarily hospitals, but also government agencies 
and manufacturers; 
Source of information in alerts: FDA, manufacturers, and subscribers. 

Type of system: Web-based customized alerts and tracking; 
Service provided: A customized Web-based recall management service 
that alerts appropriate personnel of recalls and allows them to track 
a recall's progress; 
Primary customers: Designed for hospitals but used by others including 
outpatient centers, nursing homes, and day care centers; 
Source of information in alerts: FDA, manufacturers, subscribers, U.S. 
Army Medical Material Agency, and Consumer Products Safety Commission. 

Source: GAO analysis of information from medical device recall 
stakeholders. 

[End of table] 

Stakeholders we interviewed identified several operational benefits of 
using recall notification systems. First, they indicated these systems 
allow for an increased ability to identify the universe of recalls 
rather than simply relying on receiving notices from recalling firms. 
Second, stakeholders indicated that quality assurance measures used by 
the systems help ensure that the recall notifications contain 
sufficiently detailed and accurate information. They indicated that 
personnel working for such systems will review the recall 
notifications they compile prior to sending them to the subscribers. 
If needed, those providing the recall alert services will contact the 
recalling firm and update the notice for the subscribers if 
information is unclear or missing. Third, stakeholders indicated that 
such systems can help ensure that recall notifications are routed to 
specific personnel within an institution responsible for managing the 
recall, reducing the likelihood that implementing the recall is 
delayed or overlooked. Officials from some hospitals we spoke with 
indicated that manufacturers will frequently notify the department in 
a hospital that received the product, which may not be the best point 
of contact for ensuring recalled devices are corrected or removed. 
However, one hospital system indicated that by using the more 
sophisticated alert systems, they are able to automatically forward 
recall alerts to key personnel specifically identified by the 
hospital. This ensures that only the appropriate departments at the 
hospital are alerted. Finally, stakeholders stated that these systems 
allow hospitals to identify and process recalls sooner. 

Status of UDI Initiative: 

The Food and Drug Administration Amendments Act of 2007 (FDAAA) 
required FDA to develop a UDI system--a major initiative to better 
track and identify devices.[Footnote 42] Through the UDI, FDA plans to 
require that the label of a device bear a unique identifier that is 
able to identify the device throughout its distribution and use. 
Figure 8 displays an example of the key attributes that might be 
included in a UDI. In this example, the label includes key 
information, such as a device's lot number and expiration date which 
could be scanned into databases. 

Figure 8: Example of a Potential Unique Device Identification Label: 

[Refer to PDF for image: illustration] 

The following are depicted on the label: 
Product manufacturer; 
Product number; 
Product description; 
Lot number; 
Product specification; 
Type of sterilization done to product; 
Expiration date; 
Manufacturing date; 
Expanded bar code. 

Source: Medtronic. 

[End of figure] 

FDA has been working on the UDI since 2005, prior to the enactment of 
FDAAA, and has made some preliminary decisions about the system. 
According to FDA, it currently has a proposed schedule that calls for 
implementation of the UDI in phases over several years. 

Key activities completed for the UDI include the following. 

* April 14-15, 2005: FDA held a workshop to obtain comments from 
various stakeholders on the UDI. A draft report about the UDI prepared 
by a contractor for FDA, known as "The White Paper," was provided to 
attendees prior to the workshop to use as background for workshop 
discussions. 

* August 17, 2005: The White Paper was issued and provided information 
on technologies and standards available for the UDI initiative and the 
possible benefits of automatic identification of devices. The paper 
also identified key issues FDA should consider moving forward, 
including costs of the UDI. Also, the paper incorporated stakeholder 
comments from the workshop held in April 2005.[Footnote 43] 

* March 22, 2006: Another contractor issued a report outlining the 
possible benefits of the UDI and decisions FDA must make to implement 
the system, including the technology needed to use the UDI.[Footnote 
44] 

* August 11, 2006: FDA formally solicited comments in the Federal 
Register for the UDI initiative.[Footnote 45] 

* September 27, 2007: FDAAA enacted, requiring FDA to develop and 
implement the UDI. 

* February 12, 2009: FDA held a public workshop on the UDI to identify 
remaining issues related to the establishment of a UDI system and to 
request comments on this topic. 

* November 20, 2009: FDA published the results of a pilot test of the 
UDI. The results included several recommendations for the future of 
the UDI including specific enhancements that could enhance the UDI's 
functionality.[Footnote 46] 

* November 30, 2010: Another report on pilot activities was published 
containing feedback from organizations that will label the devices and 
internal FDA stakeholders. The report stated that fewer concerns 
remain as FDA is close to releasing the UDI regulation.[Footnote 47] 

According to FDA, the UDI implementation schedule calls for a phased 
approach that will take several years to reach full-scale 
implementation. FDA is currently working on a proposed rule and 
intends to publish it and seek public comments in spring 2011, and 
issue a final rule 12 to 18 months later. According to FDA's senior 
advisor for the UDI, the proposed rule will include several key 
decisions that FDA, based on its prior studies, has reached regarding 
the UDI. These key decisions include the following. 

* Provisions for a UDI database that FDA will maintain. Manufacturers 
will send key information about their devices to FDA, which will 
maintain a database containing a device identifier for all devices 
distributed in the United States. 

* Flexibility to allow manufacturers to decide how to label their 
devices using automatic identification and data capture. This could 
mean using a linear or two-dimensional bar code, or radio frequency 
identification.[Footnote 48] 

* In addition, FDA indicated that there are other issues for which 
they have not yet made final decisions, and they are still assessing 
these before they issue a proposed rule. These include the following. 

* The labeling requirements for different devices, for example, 
riskier devices may be labeled with a unique identifier individually, 
while disposable, low-risk devices may be labeled based on how they 
are packaged (e.g., bandages will have their UDI identifier on their 
box). 

* If the UDI should have a phased implementation schedule for 
administering the identifiers, for example, class III devices--the 
most risky devices, including some that are implantable--may use the 
UDI within 1 year of publishing the final rule, while class II and 
class I devices might follow meeting the UDI requirement within 3 and 
5 years, respectively. 

Figure 9 presents a timeline of key activities since FDA began 
assessing the UDI and its planned implementation schedule. 

Figure 9: Timeline of Events for the Development of the Unique Device 
Identification Initiative: 

[Refer to PDF for image: timeline] 

Events that have occurred: 

April 14-15, 2005: 
FDA holds a meeting on the UDI to gather information from stakeholders. 

August 17, 2005: 
Report published by contractor. 

March 22, 2006: 
Second report published by contractor. 

August 11, 2006: 
FDA solicits public comments on the UDI in the Federal Register. 

September 27, 2007: 
FDAAA signed into law. 

February 12, 2009: 
FDA UDI Public Workshop held. 

November 20, 2009: 
FDA publishes results of a pilot test. 

November 30, 2010: 
Results of more FDA pilot activities published. 

Planned events: 

Spring 2011: 
FDA plans to publish proposed rule. 

2012: 
Final Rule published. 

2013: 
Class III devices required to have UDI identifier. 

2015: 
Class II devices required to have UDI identifier. 

2017: 
Class I devices required to have UDI identifier. 

Source: GAO analysis of FDA information. 

[End of figure] 

An FDA official said that the agency expects that the UDI will provide 
benefits beyond increased precision in identifying recalled devices, 
and that some benefits of the UDI will be realized immediately after 
its implementation. According to the UDI senior advisor, these 
benefits include improved tracking of adverse events associated with 
medical devices and prevention of device counterfeiting. He also 
stated that many manufacturers already use identifiers on their 
devices and should have little problem adapting to the new UDI system. 

Despite the potential benefits of the UDI, some stakeholders expressed 
concern that the success of UDIs depends on hospitals' ability to 
utilize these identifiers, and that it may be years before the 
benefits to the recall process are realized. Manufacturers we 
contacted stated that many hospitals do not use the lot and serial 
numbers currently provided by manufacturers to track devices, and FDA 
does not have authority to require providers to use the UDI. This 
concern was also reflected in comments from officials at several 
hospitals that we contacted. Some indicated that they do not have 
inventory systems in place that enable them to track devices 
throughout their hospitals. Therefore, they must manually search their 
inventory, sometimes at multiple locations. Locating a recalled device 
can be particularly difficult because a device may contain multiple 
identification numbers assigned by manufacturers and distributors for 
their own tracking purposes. Without upgrades to these hospitals' 
systems, officials acknowledged that the UDI will be less effective in 
enhancing patient safety. FDA's UDI senior advisor stated that larger 
hospitals might be more eager to adopt the technology necessary to 
track devices using the UDI once it is implemented, but acknowledged 
that benefits for the recall process are greatly dependent on 
hospitals' implementation of the UDI, which could take up to 10 years 
for many hospitals, especially smaller ones. 

[End of section] 

Appendix II: Class I Medical Device Recalls from Our Sample for Which 
Firms Were Unable to Correct or Remove All Devices: 

In some instances recalling firms are not able to correct or remove 
all of the devices subject to a recall. Of the 53 class I recalls in 
our sample of recalls that were initiated during the period of January 
1, 2005, through December 31, 2009, there were recalled medical 
devices that firms were unable to correct or remove. Table 6 includes 
information on the number of devices subject to these 21 recalls, the 
number corrected or removed, and if available, reasons firms provided 
to FDA explaining why they could not correct or remove 100 percent of 
the devices. 

Table 6: Class I Medical Device Recalls for Which Firms Were Unable to 
Correct or Remove All Devices: 

Recalled device: Centrifuge rotor; 
Reason for recall: Catastrophic failure or explosion; 
Number of devices recalled[A]: 42[B]; 
Number of devices corrected or removed[A]: 32[B]; 
Number of devices outstanding[A]: 10[B]; 
Reasons for incomplete recovery: Could not locate some users or 
devices. 

Recalled device: Endoscopic injector; 
Reason for recall: Could inject into vital organ and cause death; 
Number of devices recalled[A]: 9,984; 
Number of devices corrected or removed[A]: 4,114; 
Number of devices outstanding[A]: 5,870; 
Reasons for incomplete recovery: Users could not locate many devices, 
others were already used. 

Recalled device: Insulin syringe; 
Reason for recall: Package mislabeled, contains incorrect dose of 
insulin; 
Number of devices recalled[A]: 471,000; 
Number of devices corrected or removed[A]: 61,351; 
Number of devices outstanding[A]: 409,649; 
Reasons for incomplete recovery: Probably used by consumers or thrown 
away. 

Recalled device: Intra-aortic balloon and control system; 
Reason for recall: Fault in connector may result in incorrect pump 
volume; 
Number of devices recalled[A]: 13,570[B]; 
Number of devices corrected or removed[A]: 1,174[B]; 
Number of devices outstanding[A]: 12,396[B]; 
Reasons for incomplete recovery: Some users did not respond, other 
devices were already used. 

Recalled device: Blood glucose test strip; 
Reason for recall: Counterfeit device results in inaccurate test 
results; 
Number of devices recalled[A]: 5,292; 
Number of devices corrected or removed[A]: 1,786; 
Number of devices outstanding[A]: 3,506; 
Reasons for incomplete recovery: Presumed used or thrown away. 

Recalled device: Denervation probe; 
Reason for recall: Mislabeled as sterile; 
Number of devices recalled[A]: 539[C]; 
Number of devices corrected or removed[A]: 27[C]; 
Number of devices outstanding[A]: 512[C]; 
Reasons for incomplete recovery: Many users sterilized devices 
themselves and then used them. 

Recalled device: Heating pad; 
Reason for recall: Electrical problems, leading to fires, burns, and 
property damage; 
Number of devices recalled[A]: 408,599; 
Number of devices corrected or removed[A]: 119; 
Number of devices outstanding[A]: 408,480; 
Reasons for incomplete recovery: Firm assumed most had been disposed 
of (recall was initiated 6 years after manufacturing). 

Recalled device: Glucose monitoring system; 
Reason for recall: Displaying American units for Canadian customers; 
Number of devices recalled[A]: 33[B]; 
Number of devices corrected or removed[A]: 32[B]; 
Number of devices outstanding[A]: 1[B]; 
Reasons for incomplete recovery: Recalling firm could not locate the 
device. 

Recalled device: Infusion set; 
Reason for recall: Improper venting, leading to fluctuating insulin 
delivery; 
Number of devices recalled[A]: Unknown[D]; 
Number of devices corrected or removed[A]: 424,506; 
Number of devices outstanding[A]: Unknown[D]; 
Reasons for incomplete recovery: Single-use sets; 
impossible to know how many were available for use. 

Recalled device: Ventricular catheter; 
Reason for recall: Catheter becomes detached from base after 
implantation; 
Number of devices recalled[A]: 3,048; 
Number of devices corrected or removed[A]: Unknown[D]; 
Number of devices outstanding[A]: Unknown[D]; 
Reasons for incomplete recovery: Recalling firm did not indicate final 
disposition of all recalled devices. 

Recalled device: IV extension sets; 
Reason for recall: Manufacturing defect, possibly resulting in 
improper functioning and embolism; 
Number of devices recalled[A]: 99; 
Number of devices corrected or removed[A]: 56; 
Number of devices outstanding[A]: 43; 
Reasons for incomplete recovery: No reason given. 

Recalled device: Needle and infusion sets; 
Reason for recall: Needles "punch out," sending the core into a 
patient's body; 
Number of devices recalled[A]: Unknown[D]; 
Number of devices corrected or removed[A]: 13,988; 
Number of devices outstanding[A]: Unknown[D]; 
Reasons for incomplete recovery: Firm did not list total number of 
devices or all devices were recovered. 

Recalled device: Ventilator; 
Reason for recall: Flow valve failures; 
Number of devices recalled[A]: 277; 
Number of devices corrected or removed[A]: 276; 
Number of devices outstanding[A]: 1; 
Reasons for incomplete recovery: Missing from a nursing home (police 
report was filed). 

Recalled device: Ventilator; 
Reason for recall: Power supply failures; 
Number of devices recalled[A]: 2,270; 
Number of devices corrected or removed[A]: Unknown[D]; 
Number of devices outstanding[A]: Unknown[D]; 
Reasons for incomplete recovery: Recalling firm could not account for 
all devices and noted that some customers did not respond. 

Recalled device: Catheter; 
Reason for recall: Catheter may crack; 
Number of devices recalled[A]: 70; 
Number of devices corrected or removed[A]: 55; 
Number of devices outstanding[A]: 15; 
Reasons for incomplete recovery: Devices were used. 

Recalled device: Automatic implantable cardioverter defibrillator; 
Reason for recall: Possible short-circuiting of the device; 
Number of devices recalled[A]: 14,010; 
Number of devices corrected or removed[A]: 13,839; 
Number of devices outstanding[A]: 171; 
Reasons for incomplete recovery: Sold, but no implant record available. 

Recalled device: Pacemaker; 
Reason for recall: Device's seal may degrade allowing excess moisture 
within the pacemaker; 
Number of devices recalled[A]: 23,987; 
Number of devices corrected or removed[A]: 22,255; 
Number of devices outstanding[A]: 1,732; 
Reasons for incomplete recovery: Sold, but no implant record available. 

Recalled device: Implantable programmable pump; 
Reason for recall: Pumps may detach, interrupting drug flow; 
Number of devices recalled[A]: 1,742; 
Number of devices corrected or removed[A]: 1,497; 
Number of devices outstanding[A]: 245; 
Reasons for incomplete recovery: Some devices remain implanted or have 
an undetermined status. 

Recalled device: Implantable programmable pump; 
Reason for recall: Pump motor may stall causing drug delivery to stop 
abruptly; 
Number of devices recalled[A]: 23,895[C]; 
Number of devices corrected or removed[A]: Unknown[D]; 
Number of devices outstanding[A]: Unknown[D]; 
Reasons for incomplete recovery: Hundreds of users did not reply. 

Recalled device: Carotid stent system; 
Reason for recall: Possible detachment in the stent delivery system; 
Number of devices recalled[A]: 1,570; 
Number of devices corrected or removed[A]: 619; 
Number of devices outstanding[A]: 951; 
Reasons for incomplete recovery: Devices were either used or disposed 
at the users' facilities. 

Recalled device: Implantable defibrillator; 
Reason for recall: Potential for a short circuit and loss of device 
function; 
Number of devices recalled[A]: 6,911; 
Number of devices corrected or removed[A]: 6,787; 
Number of devices outstanding[A]: 124; 
Reasons for incomplete recovery: Sold, but no implant record available. 

Source: GAO analysis of FDA recall files. 

[A] Numbers reflect units of distribution. In some cases, devices are 
sold individually, and in other instances multiple devices are sold in 
a package (e.g., glucose strips). 

[B] Includes units distributed to Canada. 

[C] Includes worldwide units. 

[D] Documentation provided by FDA did not include these numbers. 

[End of table] 

[End of section] 

Appendix III: Comments from the Department of Health and Human 
Services: 

Department Of Health & Human Services: 
Office of the Assistant Secretary for Legislation: 
Washington, D.C. 20201: 

May 26, 2011: 

Marcia Crosse: 
Director, Health Care: 
U.S. Government Accountability Office: 
441 G Street N.W. 
Washington, DC 20548: 

Dear Ms. Crosse: 

Attached are comments on the U.S. Government Accountability Office's 
(GAO) draft report entitled, "Medical Devices: FDA Should Enhance Its 
Oversight of Recalls” (GA0-11-468). 

The Department appreciates the opportunity to review this draft report 
prior to publication. 

Sincerely, 

Signed by: 

Jim R. Esquea: 
Assistant Secretary for Legislation: 

Attachment: 

[End of letter] 

General Comments Of The Department of Health and Human Services (HHS) 
On The Government Accountability Office's (GAO) Draft Report Entitled, 
"Medical Devices: FDA Should Enhance Its Oversight Of Recalls" (GA0-11-
468) 

The Department appreciates the opportunity to review and comment on 
this draft report. 

In general, the Food and Drug Administration (FDA) believes GAO's 
draft report accurately describes the FDA's medical device recall 
program. The agency agrees with GAO's conclusions and recommendations 
and is committed to exploring each recommendation fully and to making 
appropriate implementation decisions based on the results of its 
internal analysis. 

The FDA agrees that agency recall data is a robust resource, the 
assessment of which can help mitigate health risks presented by 
defective or unsafe devices. The FDA also agrees that the agency does 
not currently conduct some of the analyses described in the GAO 
report. Although industry is responsible for analysis of design and 
manufacturing-related risks of individual medical devices, the FDA has 
an important role in recall data analysis and performs a variety of 
routine analyses to inform its regulatory activities. The FDA compiles 
recall information for every product code (medical device type), each 
medical device brand name, and every individual firm. The FDA 
aggregates recall information on related products for each 510(k) 
submission reviewed, each Establishment Inspection Report (EIR) 
review, and all adverse event analyses. The FDA's risk-based work plan 
uses recall data as one data source to identify types of medical 
devices and industry segments that should be priorities for 
inspection. Likewise, the FDA uses recall data to support its risk-
based site selection model, which identifies individual facilities 
likely to have quality system violations. Every cross-Center team 
begins with a recall analysis similar to that provided for the 
Automated External Defibrillator effort cited in the GAO report. 
Recent cross-Center teams have focused on infusion pumps and 
ventilators. 

Recommendation 1: Create a program to routinely and systematically 
assess medical device recall information, and use this information to 
proactively identify strategies for mitigating health risks presented 
by defective or unsafe devices. This assessment should be designed, at 
a minimum, to identify trends in the numbers and types of recalls, 
devices most frequently being recalled, and underlying causes of 
recalls. 

The FDA agrees it should assess medical device recall information more 
routinely and systematically, and that it should use this information 
to mitigate health risks. The FDA plans to perform a systematic 
evaluation of potential recall analyses. The agency plans to consider 
recall processing times, recalled device types and medical 
specialties, types of device failures observed, quality of recall 
notifications, and other elements that elevate the risk to public 
health. Based on the evaluation, the agency will develop more routine 
analysis and reporting. In the meantime, the FDA plans to continue to 
analyze numbers and types of recalls on a yearly basis. 

Recommendation 2: Clarify procedures for conducting medical device 
recall audit checks to improve the ability of investigators to perform 
these checks in a consistent manner. 

The FDA generally agrees it should update and clarify its recall audit 
check procedures and is committed to reevaluating its procedures for 
conducting recall audit checks. In recent years, the agency has 
experienced significant increases in the number of recalls of FDA-
regulated products, some of which involve the largest and most complex 
recalls in FDA history. To address these issues and to better 
understand the current recall business processes, the FDA has 
undertaken a recall process improvement study that is intended to 
examine the overall recall process in protecting the public health and 
removing violative products from the marketplace. This study was 
recently completed, and the FDA plans to convene a working group to 
evaluate the results and to develop a strategy for the implementation 
of the recommendations. Furthermore, the FDA will continually review 
its recall procedures to determine whether specific additions, 
updates, and clarifications may be needed to the recall program, 
including the audit check process. 

In addition, the FDA has already begun implementing some initiatives 
for improving the recall audit check process, which currently is 
manual, for all FDA-regulated products. For example, and as noted in 
GAO's report, the FDA has invested in automating its current recall 
audit check process, which will allow the agency to automate the 
creation, issuance, assigning, and tracking of recall audit check 
assignments, as well as to establish an electronic mechanism for 
reporting or documenting recall audit check results. 

Recommendation 3: Develop explicit criteria for assessing whether 
recalling firms have performed an effective correction or removal 
action. 

The FDA generally agrees it should clarify its criteria for assessing 
whether recalling firms have performed an effective correction or 
removal action. The FDA acknowledges it needs to update its process 
and procedures, and recognizes that standardized guidance will 
strengthen the management of the recall process and promote a 
consistent approach for assessing recalling firms' efforts and 
effectiveness in correcting or removing violative products from the 
marketplace. The FDA is committed to continually analyzing the 
procedures and the criteria it uses to determine when recalling firms 
have performed an effective correction or removal action. The FDA 
working group which will be tasked with evaluating the results of the 
recently completed recall process improvement study will use the study 
results to develop an informed strategy for implementing this 
recommendation. 

Recommendation 4: Document the agency's basis for terminating 
individual recalls. 

The FDA generally agrees that the agency should document its 
termination decisions and is committed to reevaluating its recall 
termination procedures and documentation. The FDA working group that 
will be tasked with evaluating the results of the recently-completed 
recall process improvement study will consider the study results to 
make recommendations to FDA management on the need for specific 
additions, updates, and clarifications to the recall termination 
process. 

[End of section] 

Appendix IV: GAO Contact and Staff Acknowledgments: 

GAO Contact: 

Marcia Crosse, (202) 512-7114 or crossem@gao.gov: 

Staff Acknowledgments: 

In addition to the contact named above, Geri Redican-Bigott, Assistant 
Director; Kaycee Glavich; Cathleen Hamann; Eagan Kemp; Julian Klazkin; 
Zachary Levinson; David Lichtenfeld; Daniel Ries; Christina C. Serna; 
and Katherine Wunderink made key contributions to this report. 

[End of section] 

Footnotes: 

[1] 21 U.S.C. § 360h(e), 21 C.F.R. pt. 810 (2010). The steps that FDA 
follows for such mandatory recalls are separate and distinct from 
those followed during a voluntary recall. The steps that FDA follows 
during voluntary recalls are described at subpart C, part 7, of title 
21 of the Code of Federal Regulations. 

[2] For purposes of this report, we use the term guidance to refer to 
FDA's advice or recommendations to recalling firms managing voluntary 
recalls and the term procedures to refer to FDA's instructions to 
agency staff overseeing these recalls. 

[3] GAO, Medical Devices: FDA Can Improve Oversight of Tracking and 
Recall Systems, [hyperlink, 
http://www.gao.gov/products/GAO/HEHS-98-211] (Washington, D.C.: Sept. 
24, 1998). 

[4] GAO, Medical Devices: FDA's Premarket Review and Postmarket Safety 
Efforts, [hyperlink, http://www.gao.gov/products/GAO-11-556T] 
(Washington, D.C.: Apr. 13, 2011). 

[5] While FDA has authority to order a mandatory recall, it did not 
exercise this authority during the period we reviewed. See 21 U.S.C. § 
360h(e), 21 C.F.R. pt. 810 (2010). Also, our information does not 
include devices that a firm may have voluntarily taken off the market 
for other, less serious, reasons. For example, a market withdrawal is 
a firm's correction or removal of a distributed device that involves 
no violation or a minor violation of the laws FDA administers and for 
which FDA would not initiate legal action. 21 C.F.R. § 806.2(h) 
(2010). A stock recovery is a firm's correction or removal of a device 
that has not been marketed or that has not left the direct control of 
the manufacturer. 21 C.F.R § 806.2(l) (2010). FDA does not consider 
market withdrawals and stock recoveries to be recalls. 

[6] FDA tracks recalls by both the recall event, that is, the actual 
process of implementing a recall, and the products being recalled. A 
single recall event could involve multiple products (e.g., different 
sizes and models of the same device). We use the term recall in the 
report to refer to recall events. 

[7] The status of the recalls is as of April 16, 2010, the date FDA 
provided us with the extract from RES. 

[8] See GAO, Standards for Internal Control in the Federal Government, 
[hyperlink, http://www.gao.gov/products/GAO/AIMD-00-21.3.1] 
(Washington, D.C.: Nov. 1999). 

[9] A small percentage of devices enter the market by other means, 
such as through the humanitarian device exemption process that allows 
market entry, without adherence to certain requirements, for devices 
benefiting patients with rare diseases or conditions. See 21 U.S.C. § 
360j(m), 21 C.F.R. pt. 814, subpart H (2010). In addition, many other 
less risky types of class I and II devices are also exempt from FDA's 
premarket review. 

[10] Substantial equivalence or substantially equivalent means that 
the device has the same intended use as another legally marketed 
device and the same technological characteristics, or that the device 
has different technological characteristics and information submitted 
to FDA demonstrates that the device is as safe and effective as the 
legally marketed device and does not raise different questions of 
safety or effectiveness. 21 U.S.C. § 360c(i)(1)(A). 

[11] According to FDA, performance testing results should be submitted 
if there are important descriptive differences between the device and 
other devices of the same type or if the descriptive characteristics 
for the new device are not precise enough to ensure comparability. In 
these instances, the most appropriate bench testing, animal testing, 
or both to address the performance issue should be provided, and 
summary information on the testing should generally suffice. 

[12] From 2003 through 2007, 79 percent of the riskiest medical 
devices (class III devices) were approved through the PMA process. As 
we reported in 2009, FDA has cleared some class III devices through 
the 510(k) process, rather than approving the devices through the more 
stringent PMA process. See GAO, Medical Devices: FDA Should Take Steps 
to Ensure That High-Risk Device Types Are Approved Through the Most 
Stringent Premarket Review Process, GAO-09-190 (Washington, D.C.: Jan. 
15, 2009). 

[13] 21 C.F.R. § 7.3(g) (2010). A removal is the physical removal of a 
device from its point of use to some other location for repair, 
modification, adjustment, relabeling, destruction, or inspection. A 
correction may involve these actions without the physical removal of a 
device from its point of use. 21 C.F.R. § 806.2(d, i) (2010). 

[14] Stakeholders such as distributors and device users, and anyone 
else who received, purchased, or used the product being recalled are 
referred to as "consignees." 21 C.F.R. §§ 7.3(n), 806.2(c) (2010). In 
this report, we generally refer to these entities as customers and 
device users. 

[15] A given recall may involve a single device or multiple devices. 
For example, a recall may include different sizes of the same device, 
or a device made up of multiple components (i.e., screws, bolts, 
tubes). 

[16] The firm will contact one of FDA's district offices depending 
upon the location from which it chooses to manage the recall. This 
district will have primary responsibility for monitoring the recall. 
Each district has a recall coordinator, who among other duties, 
processes medical device recalls and monitors the progress of the 
firm's actions. 

[17] 21 C.F.R. § 806.10(a), (b) (2010). Reports are not required for 
recalls involving problematic devices that do not pose a health risk. 

[18] FDA officials indicated that although this time frame was in 
effect at the time of our review, they have recently adjusted 
classification time frames, which now range from 26 to 40 days 
depending on the nature of the recall. 

[19] See 21 C.F.R. § 7.3(m) (2010). 

[20] Audit checks, which generally cover from 2 to 10 percent of the 
total number of distributors and device users, are typically conducted 
by FDA on class I and class II recalls. Audit checks are generally not 
conducted on class III recalls. Audit checks are separate from 
effectiveness checks, which are conducted by the recalling firm. 

[21] A firm's completion of a recall does not necessarily indicate 
that it corrected or removed 100 percent of the recalled devices, but 
rather that it made reasonable efforts to contact those affected by 
the recall and obtained as many devices as possible. 

[22] General hospital and personal use devices include bandages, 
examination gowns, infusion pumps, and stretchers. Diagnostic 
chemistry devices include test systems, such as those for blood 
glucose. 

[23] Information on mode of market entry--PMA, 510(k), or exempt--was 
missing for 737 (21 percent) of the 3,510 recalls. Our analysis is 
therefore based on the available data related to 2,773 recalls. 

[24] This includes recalls involving multiple devices (such as kits 
containing a pump and a catheter) some of which were cleared through 
the 510(k) process, and others approved through the PMA process. It 
also includes devices that are exempt from FDA review, such as most 
class I devices. 

[25] Information on mode of market entry--PMA, 510(k), or exempt--was 
missing for 30 (23 percent) of the 131 class I recalls that occurred 
during calendar years 2005 through 2009. Therefore, our above analysis 
is based on the available data related to 101 recalls. 

[26] Two separate studies also analyzed class I recalls from 2005 to 
2009 and found similar statistics as ours. See D. Zukerman, Paul 
Brown, Steven Nissen, Medical Device Recalls and the FDA Approval 
Process. Archives of Internal Medicine (Feb. 14, 2011) and related 
testimony at U.S. House of Representatives, Energy and Commerce 
Subcommittee on Health Hearing on "Impact of Medical Device Regulation 
on Jobs and Patients" (Feb. 17, 2011). See also R. Hall, Using Recall 
Data to Assess the 510(k) Process. Presentation at Institutes of 
Medicine Meeting (July 28, 2010) and related testimony at U.S. House 
of Representatives, Energy and Commerce Subcommittee on Health, 
Hearing on "Impact of Medical Device Regulation on Jobs and Patients" 
(Feb. 17, 2011). 

[27] Of the 2,050 medical device recalls that had been terminated, 22 
were missing valid termination dates and hence were not included in 
our analysis. 

[28] According to FDA, in some cases a firm initiates a recall without 
contacting FDA, thus the agency may not learn about the recall until a 
much later date. 

[29] We previously reported that FDA failed to meet its 90-day 
termination guideline in 33 percent of the 36 recalls we reviewed. See 
[hyperlink, http://www.gao.gov/products/GAO/HEHS-98-211]. 

[30] We previously reported on the importance of establishing and 
using metrics as a management tool. See, for example, GAO, Food and 
Drug Administration: Opportunities Exist to Better Address Management 
Challenges, [hyperlink, http://www.gao.gov/products/GAO-10-279] 
(Washington, D.C.: Feb. 19, 2010). 

[31] We previously reported that FDA faced challenges fulfilling and 
managing its growing medical product oversight responsibilities and 
should develop complete estimates of its resource needs. See GAO, Food 
and Drug Administration: FDA Faces Challenges Meeting Its Growing 
Medical Product Responsibilities and Should Develop Complete Estimates 
of Its Resource Needs, [hyperlink, 
http://www.gao.gov/products/GAO-09-581] (Washington, D.C.: June 19, 
2009). 

[32] Statement of Dr. William Maisel, CDRH Deputy Director for 
Science, at FDA Public Workshop on External Defibrillators 
(Washington, D.C.: Dec. 15-16, 2010). 

[33] According to FDA, this initiative also includes plans to develop 
strategies for improving CDRH's processes for classifying recalls and 
notifying the public about recalls. 

[34] For 4 of the 53 recalls we reviewed, FDA identified the problem 
that led to the recalls through postmarket establishment inspections, 
unrelated to recalls. Thus, FDA did not need to conduct initial recall 
inspections. 

[35] FDA's guidance at the time of our review recommended that CDRH 
classify a recall within 31 calendar days of obtaining information 
needed to classify the recall. 

[36] We obtained the current status of these recalls through our 
reviews of the recall files and discussions with FDA district office 
officials. These discussions took place between December 2010 and 
February 2011. 

[37] FDA's procedures note that audit checks should be conducted for 
all class I recalls. In 8 of the 53 recalls FDA did not conduct audit 
checks. In six of these cases, the recall file contained written 
documentation explaining why audit checks were not conducted. 

[38] For 2 of the 45 recalls there was evidence that FDA conducted 
audit checks, but FDA did not provide copies of the audit check forms. 
Thus, the 2,196 audit checks forms we reviewed were for a total of 43 
recalls. 

[39] For two recalls, FDA was unable to locate the audit check forms. 
Thus, we have excluded these recalls from the calculation of the 
average number of audit checks. 

[40] See [hyperlink, http://www.gao.gov/products/GAO/AIMD-00-21.3.1]. 

[41] Published studies we reviewed include: Eastern Research Group, 
Inc., ERG Final Report: Unique Identification for Medical Devices, a 
report prepared for FDA, March 2006, and Susan J. Griffey, Andy 
Martin, Marjorie Odle, and M.J. Wylie, SSS-GHX Development of the 
Prototype Unique Device Identifier Database: Report of a Pilot Test on 
Usability and Feasibility, a report prepared for FDA, (November 2009). 

[42] Pub. L. No. 110-85, § 226, 121 Stat. 823, 854. 

[43] The ECRI Institute, Task 4 White Paper - Automatic Identification 
of Medical Devices - Final Version, a report prepared for FDA, August 
2005. 

[44] Eastern Research Group, Inc., ERG Final Report: Unique 
Identification for Medical Devices, a report prepared for FDA, March 
2006. 

[45] Docket No. FDA-2008-N-0661, CDRH 200866. Unique Device 
Identification System; Public Workshop; Request for Comments. 

[46] Susan J. Griffey, Andy Martin, Marjorie Odle, MJ Wylie, SSS-GHX 
Development of the Prototype Unique Device Identifier Database: Report 
of a Pilot Test on Usability and Feasibility, a report prepared for 
FDA, (November 2009). 

[47] FDA, Results of FDA Pilot Activities To Explore Opportunities and 
Challenges With the Implementation of a Unique Device Identifier 
System (Washington, D.C., 2010). 

[48] A two-dimensional barcode stores information horizontally and 
vertically allowing the two-dimensional barcodes to store more 
information than linear bar codes, which store information only 
horizontally. Radio frequency identification allows for information 
about a device to be transmitted between a tag on the device and a 
reader using radio waves. 

[End of section] 

GAO's Mission: 

The Government Accountability Office, the audit, evaluation and 
investigative arm of Congress, exists to support Congress in meeting 
its constitutional responsibilities and to help improve the performance 
and accountability of the federal government for the American people. 
GAO examines the use of public funds; evaluates federal programs and 
policies; and provides analyses, recommendations, and other assistance 
to help Congress make informed oversight, policy, and funding 
decisions. GAO's commitment to good government is reflected in its core 
values of accountability, integrity, and reliability. 

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through GAO's Web site [hyperlink, http://www.gao.gov]. Each 
weekday, GAO posts newly released reports, testimony, and 
correspondence on its Web site. To have GAO e-mail you a list of newly 
posted products every afternoon, go to [hyperlink, http://www.gao.gov] 
and select "E-mail Updates." 

Order by Phone: 

The price of each GAO publication reflects GAO’s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO’s Web site, 
[hyperlink, http://www.gao.gov/ordering.htm]. 

Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537. 

Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional 
information. 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 

Web site: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]: 
E-mail: fraudnet@gao.gov: 
Automated answering system: (800) 424-5454 or (202) 512-7470: 

Congressional Relations: 

Ralph Dawn, Managing Director, dawnr@gao.gov: 
(202) 512-4400: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7125: 
Washington, D.C. 20548: 

Public Affairs: 

Chuck Young, Managing Director, youngc1@gao.gov: 
(202) 512-4800: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7149: 
Washington, D.C. 20548: