Fire Safety: Comprehensive Information on Fire Incidences in	 
Federal Facilities Is Lacking (20-AUG-01, GAO-01-879).		 
								 
Developing standards that protect against fire and testing	 
products against those standards are critical in promoting fire  
safety. According to the National Fire Protection Association,	 
business office properties, including federal offices and other  
federal civilian facilities, annually experience thousands of	 
fires, over $100 million in property losses, and dozens of	 
casualties each year. Fires, even relatively small ones, can have
tragic and costly consequences. Knowing the numbers and types of 
fires in the workplace, as well as the causes of fires and any	 
products involved, is critical for understanding the extent of	 
the risk of fire and can lead to an identification and		 
implementation of steps to reduce this risk. Some private-sector 
organizations track the number of fires in different types of	 
facilities and their causes. Such information is used to manage  
this risk and reduce property damage, injuries, and the loss of  
life. However, the federal government collects very little	 
information on fires and lacks information on the risk of fire in
its facilities. Without this information, the government cannot  
provide timely information on the causes of fires in federal	 
facilities to standards-development organizations for their use  
in developing and revising standards, testing procedures, and	 
certification decisions. Collecting and analyzing data on the	 
risk of fire in its facilities could enable the government to	 
better protect its employees and enhance its ability to 	 
participate in producing standards that would better protect the 
public at large from fire.					 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-01-879 					        
    ACCNO:   A01382						        
    TITLE:   Fire Safety: Comprehensive Information on Fire Incidences
             in Federal Facilities Is Lacking                                 
     DATE:   08/20/2001 
  SUBJECT:   Property losses					 
	     Safety regulation					 
	     Safety standards					 
	     Data collection					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-01-879
     
Report to the Honorable Sherwood Boehlert, Chairman, Committee on Science,
House of Representatives

United States General Accounting Office

GAO

August 2001 FIRE SAFETY Comprehensive Information on Fire Incidences in
Federal Facilities Is Lacking

GAO- 01- 879

Page i GAO- 01- 879 Fire Safety Letter 1

Results in Brief 3 Background 4 Federal Government Relies Heavily on
Consensus- Based Standards

and Testing 6 Comprehensive Data on Fire Incidences in Federal Facilities
Are

Lacking 7 Conclusion 14 Recommendation for Executive Action 15 Agency
Comments and Our Evaluation 15

Appendix I Objectives, Scope, and Methodology 18

Appendix II Sources Contacted by GAO 22

Tables

Table 1: Known Fires in Which Omega Fire Sprinklers Failed to Activate 12

Abbreviations

ANSI American National Standards Institute CFOI Census of Fatal Occupational
Injuries GSA General Services Administration NFPA National Fire Protection
Association NFIRS National Fire Incident Reporting System OMB Office of
Management and Budget OSHA Occupational Safety and Health Administration
Contents

Page 1 GAO- 01- 879 Fire Safety

August 20, 2001 The Honorable Sherwood Boehlert Chairman, Committee on
Science House of Representatives

Dear Mr. Chairman: Developing standards that protect against fire and
testing products against those standards are critical in promoting fire
safety. According to the National Fire Protection Association (NFPA),
business office properties, including federal offices and other federal
civilian facilities, annually experience thousands of fires, over $100
million in property losses, and dozens of casualties each year. The
government and the public rely on product standards, testing, and
certification for protection from fires. For example, organizations such as
Underwriters Laboratories, the American Society for Testing and Materials,
and NFPA are ?standards- development organizations? that are part of the
private sector process for developing standards on a voluntary, consensus,
and largely self- regulated basis. Generally, the technical committees of
these organizations include manufacturers, government officials, consumer
representatives, and others who discuss and propose standards and testing
procedures.

The nation?s system for developing standards and testing products to certify
their compliance with those standards is complex. 1 The system consists of a
decentralized, largely self- regulated network of private independent
standards- development organizations, testing laboratories, and government
agencies. For example, there are about 50,000 private sector voluntary
standards, developed by more than 620 organizations. This number does not
include over 44,000 regulatory and procurement standards developed by some
80 federal regulatory and procurement authorities, or other codes, rules,
and regulations containing standards that have been adopted by state and
local governments.

The American National Standards Institute (ANSI) is a private nonprofit
organization of codes and standards developers and other organizations.

1 A standard is a prescribed set of rules, conditions, or requirements
concerning definitions of terms; classification of components; specification
of materials, performance, or operations; delineation of procedures; or
measurement of quantity or quality in describing materials, products,
services, or practices.

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 01- 879 Fire Safety

ANSI establishes rules for developing standards in an open, inclusive, and
transparent way, on the basis of the consensus of the parties represented in
the technical committees. Federal agencies, like the Consumer Product Safety
Commission, the Department of Labor?s Occupational Safety and Health
Administration (OSHA), and the General Services Administration (GSA), use
standards established through this consensus method and the results of
testing that determines product compliance with these standards to help
ensure the quality and safety of the goods and services that they and the
public purchase. Without standards that are developed through the consensus
process to accomplish this purpose, federal agencies would have to
independently develop their own standards to evaluate the safety of millions
of products.

Federal agencies, along with state and local government, have long
maintained responsibilities for identifying and responding to the national
fire threat. Most significantly, the U. S. Fire Administration- an agency
within the Federal Emergency Management Agency- has been statutorily
designated as the lead federal agency for coordinating fire data collection
and analysis. It also maintains a national fire incident database and
provides training and technical assistance to local fire departments and
other local enforcement agencies. Also, the Department of Commerce?s
National Institute of Standards and Technology performs research on
firerelated issues.

In a letter to us and in subsequent meetings, you expressed interest in the
adequacy of standards that promote fire safety and associated procedures for
testing products in protecting federal employees and the public. You were
concerned that some products, which had been tested and certified as meeting
these standards, had nonetheless failed in the market but were still being
used in federal facilities. For example, in the early 1990s, millions of a
specific type of sprinkler head commonly installed in federal facilities
failed to activate in fires. In October 1998 the Consumer Product Safety
Commission recalled them.

In response to your concerns and discussions with your office, this report 
provides information on the federal government?s reliance on private

sector voluntary fire standards and testing products against those standards
and  discusses whether data that are available about fire incidents and
their

causes in civilian federal facilities are sufficient to allow federal
agencies to determine if they need to take action to protect federal workers
from the threat of fire.

Page 3 GAO- 01- 879 Fire Safety

To provide this information, we identified and reviewed existing standards-
development procedures, testing protocols, and certification programs. We
also interviewed representatives of major standardsdevelopment
organizations, testing laboratories, and certification organizations. We
also interviewed fire experts from federal agencies and from major fire
safety- related associations and fire departments. We also obtained data
from the available federal and NFPA fire- incident databases and reporting
systems. Finally, as requested, we provide information on the failure of
Omega sprinkler heads to activate and on concerns about the flammability of
information technology equipment. Appendix I provides a detailed discussion
of our scope and methodology. Appendix II provides a list of the groups we
contacted for this review.

When erecting facilities; renovating offices; and purchasing equipment,
materials, and supplies, federal agencies rely on the fire safety standards
promulgated by standards- development organizations and testing to those
standards by private, independent laboratories and product certification
organizations. For example, GSA- the federal agency responsible for listing
goods and services that are approved for procurement by federal agencies-
uses product certifications as a way of ensuring the quality and safety of
these goods and services. In addition, federal agencies often incorporate
those standards into regulations, such as the Consumer Product Safety
Commission?s consumer product safety regulations. Such reliance is widely
accepted nationwide and internationally, and the federal government has long
encouraged its agencies to rely on private sector, consensus- based
standards. In 1983, an Office of Management and Budget (OMB) circular
encouraged agencies to use these standards to the maximum extent
practicable; and in March 1996, the Congress enacted the National Technology
Transfer and Advancement Act of 1995 (Public Law 104- 113), requiring
agencies to use these standards, except when it is inconsistent with
applicable law or otherwise impractical. The circular and the act also
direct federal agencies to consult with and participate, when appropriate,
in standards- development organizations and report to OMB when they do not
use private sector, voluntary standards in their procurement or regulatory
activities.

The federal government has no comprehensive, centralized database regarding
the incidence of fires in federal facilities or the causes of such fires. As
reported by NFPA, from 1993 through 1997, fires in office facilities,
including federal civilian facilities, annually caused about 90 injuries and
about $130 million in property damage. To address this fire risk, certain
parts of the private sector- for example, a major hotel chain Results in
Brief

Page 4 GAO- 01- 879 Fire Safety

and some insurance organizations- track the number of fires in different
types of facilities and their causes. Although responsible for serving as
the lead federal agency for coordinating fire data collection and analysis
by maintaining a national fire incident database, the U. S. Fire
Administration does not collect data on the number of fires in federal
office facilities, the causes of those fires, or the specific types of
products involved in fires. According to its comments on a draft of our
report, the Fire Administration does not have the resources or authority to
implement a nationwide study of fires in federal workspace. GSA does not
systematically collect information on fires that have occurred in the
facilities for which it is responsible- about 330 million square feet in
over 8,300 buildings. In addition, the nonfederal NFPA does not gather
specific information about whether a fire occurred on private or government
property or whether the fire involved specific products. As a result of a
lack of centralized data collection and reporting systems, relatively little
assurance exists that the government has sufficient knowledge of the number
and causes of fires in federal facilities to take appropriate action to
protect federal employees from the threat of fire. For example, the Omega
fire sprinkler system that failed in numerous locations and in laboratory
tests as early as 1990 was not recalled until 1998 and has only recently
been replaced at some major government facilities, such as the Smithsonian
and the Library of Congress. This report contains a recommendation aimed at
determining whether a systematic collection of data on fires in federal
workspace would provide information useful to the federal government in its
efforts to reduce the risk of fires.

We provided a draft of our report to the U. S. Fire Administration, GSA, the
Department of Labor, and the Consumer Product Safety Commission. The
Director of the Fire Administration?s National Fire Data Center agreed, in
principle, with our recommendation. GSA senior program officials suggested
the deletion of a statement that it could not provide us with complete
information on fires that caused over $100,000 damage in federal facilities,
but we declined to make this change because the statement was germane to our
discussion. GSA also provided additional information, which we incorporated
into our report. The Department of Labor and the Consumer Product Safety
Commission provided technical and editorial comments, which we incorporated
where appropriate. Because of its role in testing Omega sprinklers, we sent
a copy of the draft to Underwriters Laboratories, which had no comments.

In the United States, product safety, including fire safety, is largely
promoted through a process of consensus- based standards and voluntary
Background

Page 5 GAO- 01- 879 Fire Safety

certification programs. ANSI establishes requirements to ensure that
standards are formulated through a consensus- based process that is open and
transparent and that adequately considers and resolves comments received
from manufacturers, the fire safety community, consumers, government
agencies, and other stakeholders. Standards are generally developed in the
technical committees of organizations that include independent laboratories,
such as Underwriters Laboratories; and trade and professional associations,
such as the American Society for Testing and Materials. These entities form
a decentralized, largely self- regulated network of private, independent,
standards- development organizations. For those organizations that choose to
follow ANSI procedures, ANSI performs audits and investigations to ensure
that standards- development organizations follow approved consensus- based
procedures for establishing standards. Standards promulgated by such
organizations can become part of a system of American National Standards
currently listed by ANSI. Overall, according to NFPA, the U. S. standards
community maintains over 94,000 active standards, both American National
Standards and others. These 94,000 active standards include private sector
voluntary standards as well as regulatory and procurement standards.

The process of developing consensus- based standards is designed to balance
the needs of consumers, federal and nonfederal regulators, and
manufacturers. According to ANSI officials, new standards are commonly
adopted or existing ones are frequently revised because manufacturers
express a need for such actions on the basis of the development of new
products. Representatives of other parties- such as regulators or consumers-
may raise concerns about product safety and performance.

For marketing and consumer safety purposes, product manufacturers may have
their products tested at independent testing laboratories to certify that
the products meet applicable product standards. This testing and
certification process is called ?product conformity testing and
certification.? Some local, state and federal agencies require such testing
and certification. For example, manufacturers of electrical home appliances
have their products tested and certified by Underwriters Laboratories to
enable them to attest that the products meet safety standards regarding
fire, electrical shock, and casualty hazards. Alternatively, where
acceptable, manufacturers can certify on their own that their products were
tested and met applicable standards.

Standards are also voluntarily accepted and widely used by manufacturers and
regulatory agencies to provide guidance and specifications to manufacturers,
contractors, and procurement officials. Each year millions

Page 6 GAO- 01- 879 Fire Safety

of products are sold in the United States and throughout the world that bear
the mark of testing organizations. Consumers, manufacturers, and federal
agencies follow the very widespread, internationally recognized practice of
relying on consensus standards and testing at laboratories to promote public
safety. In the case of facilities and residences, the most extensive use of
the standards is their adoption into model building codes by reference.
Model building codes contain standards published by many organizations,
including professional engineering societies, building materials trade
associations, federal agencies, and testing laboratories.

When erecting facilities; renovating offices; and purchasing equipment,
materials, and supplies, federal agencies rely on the fire safety standards
developed by private standards- development organizations. Furthermore, the
federal government has historically encouraged its agencies to use standards
developed by these organizations. For example, in its 1983 Circular A- 119,
OMB encouraged agencies to use these standards. Moreover, the National
Technology Transfer and Advancement Act of 1995 requires agencies to use
standards developed or adopted by voluntary consensus bodies, except when it
is inconsistent with applicable law or otherwise impractical. Essentially,
OMB Circular A- 119 and the act direct federal agencies to use voluntary
consensus standards whenever possible. They also direct federal agencies to
consult with and participate, when appropriate, in standards- setting
organizations and provide explanations when they do not use voluntary
consensus standards in their procurement or regulatory activities. As of
June 2001, according to NFPA, about 15 percent of the estimated 94,000
standards effective in the United States had been developed by civilian
federal agencies. Furthermore, the Public Buildings Amendments of 1988
require GSA to construct or alter buildings in compliance with the national
building codes and other nationally recognized codes to the maximum extent
feasible.

Federal agencies also engage in a variety of activities related to
certifying that products conform to standards. For example, the National
Institute of Standards and Technology publishes directories listing more
than 200 federal government procurement and regulatory programs in which
agencies are actively involved in procuring or requiring others to procure
products meeting certification, accreditation, listing, or registration
requirements. Furthermore, many federal agencies participate in the
development of fire standards and product- testing procedures. For example,
GSA participates on technical committees, such as those of NFPA and
Underwriters Laboratories. As a result, GSA specifies numerous products and
building code regulations that meet standards and testing Federal Government

Relies Heavily on Consensus- Based Standards and Testing

Page 7 GAO- 01- 879 Fire Safety

requirements from standards- development organizations and testing
laboratories. In addition, voluntary standards and the testing of products
to those standards are widely accepted by other civilian federal agencies,
such as the departments of Agriculture, Housing and Urban Development, the
Interior, Labor, Transportation, and the Treasury as well as the
Environmental Protection Agency.

The federal government has no comprehensive, centralized database regarding
the incidence of fires in federal facilities or the causes of such fires.
According to NFPA, fires in office facilities, including federal civilian
facilities, annually cause about 90 injuries and about $130 million in
property damages. Although responsible for maintaining a national fire
incident database and for serving as the lead agency in coordinating fire
data collection and analysis, the U. S. Fire Administration does not collect
data on the number of fires in federal office facilities and the causes of
those fires, nor about specific types of products involved in the fires. For
its part, GSA collects a minimal amount of information in the facilities for
which it is responsible- about 330 million square feet in over 8,300
buildings- to determine the number and causes of fires that have occurred in
the facilities. In addition, like the U. S. Fire Administration, NFPA does
not gather specific information about whether a fire occurred on private or
government property or whether the fire involved specific products. Thus,
these databases do not contain sufficiently detailed data to allow the
identification of fire incidents in federal facilities or fires associated
with specific product defects. Also, the government does not have a
mechanism for providing fire incident data to standardsdevelopment
organizations when they consider the revision of product standards and
testing procedures. As a result of a lack of detailed data collection and
reporting systems, the government cannot assess the number and causes of
fires in federal facilities and therefore cannot determine if any action is
needed to ease the threat of fire.

Certain private sector firms take steps to identify the nature of the fire
threat in their facilities. For example, to help insurance companies,
communities, and others evaluate fire risks, the Insurance Services Office,
an affiliate of the insurance industry of the United States, maintains
detailed records and performs investigations about individual properties and
communities around the country, including such factors as the physical
features of buildings, detailed engineering analyses of building
construction, occupancy hazards, and internal and external fire protection.
In addition, the Marriott Corporation, a worldwide hotel chain, maintains
Comprehensive Data

on Fire Incidences in Federal Facilities Are Lacking

Data Regarding Fires in Federal Facilities Do Not Provide Sufficient
Information for Determining if Additional Actions Are Needed to Protect
Federal Employees

Page 8 GAO- 01- 879 Fire Safety

data on fires throughout its facilities. According to a Marriott official,
Marriott uses this information to assess the risk of fire in its facilities
and to take corrective actions.

At the same time, the number and causes of fires in federal workspace are
not known. The federal government- an employer of over two million civilian
employees- does not have a system for centrally and comprehensively
reporting fire incidents in its facilities and the causes of those
incidents. For example, according to GSA officials, the agency- which
manages over 300 million square feet of office space-- collects information
on fires that cause over $100,000 in damage. However, when we requested this
information, GSA could not provide it and provided examples of only two
fires. According to a GSA official, GSA cancelled a requirement for its
regional offices to report smaller fires to a central repository. GSA
explained that it found the task of reporting smaller fires to be very labor
intensive and time consuming. GSA also found that analysis of the reported
information could not determine specific fire trends.

Databases that are available and maintained by federal agencies- such as
databases of the Department of Labor, Consumer Product Safety Commission,
and U. S. Fire Administration- do not provide sufficient detail for
determining the number and causes of fires in federal facilities, including
the products involved in the fires. For example, according to the Department
of Labor (Labor), 7 civilian federal employees died (excluding the 21 who
died in forest or brush fires), and 1,818 civilian federal employees were
injured while at work as a result of fires or explosions between 1992 and
1999. 2 Although Labor gathers information about federal employees? injuries
and fatalities caused by fires, this information does not identify details,
such as the cause of the fire. Furthermore, because of a lack of reporting
detail, the data do not lend themselves to an analysis of what specific
products may have been involved in the fire and whether the product had been
certified as meeting appropriate product standards.

2 These fatalities do not include federal employee deaths due to bombings,
such as the Oklahoma City bombing or incidents overseas, such as the August
1998 bombing of the US Embassy in Dar Es Salaam, Tanzania. However, the
number of civilian injuries among federal employees does include bombing
victims. For example, according to the Department of Labor, the number of
injuries among civilian federal employees includes many injuries that
resulted from the 1995 bombing of the Murrah Federal Building in Oklahoma
City. The Department of Labor could not provide additional detail regarding
injuries due to bombings.

Page 9 GAO- 01- 879 Fire Safety

Within Labor, OSHA?s Office of Federal Agency Programs, the Bureau of Labor
Statistics, and the Office of Workers? Compensation Programs routinely
gather information about federal employee injuries and fatalities. OSHA?s
Office of Federal Agency Programs, whose mission is to provide guidance to
each federal agency on occupational and health issues, also collects annual
injury statistics from each federal agency. These statistics are in
aggregated form, however, and do not provide detail about the nature or
source of the injury.

The Department of Labor?s Bureau of Labor Statistics has been collecting
information on federal employee fatalities since 1992 through its Census of
Fatal Occupational Injuries (CFOI). This census contains information
regarding work- related fatality data that the federal government and the
states have gathered from workers? compensation reports, death certificates,
the news media, and other sources. According to the CFOI, between 1992 and
1999, 7 civilian federal employees were fatally injured due to fire- related
incidents while working (excluding the 21 who died in brush or forest
fires). Although the fatal injuries census does identify federal employee
fatalities due to fires, it does not contain details about the fire, such as
the cause of the fire or the types of products or materials that may have
been involved in the fire.

Also within the Department of Labor, the Office of Workers? Compensation
Programs maintains information about federal employees or families of
federal employees who have filed claims due to work- related traumas. The
office was able to provide from its database information about the claims of
federal employees or their families resulting from firerelated incidents.
According to the Office of Workers? Compensation, between 1992 and 1999
1,818 civilian federal employees were injured in federal workspace as a
result of fire- related incidents while working. However, this information
includes data only for those federal employees who actually filed claims.
Similar to CFOI data, this database does not contain additional details
about the fire, such as the cause of the fire or the types of products or
materials that may have been involved in the fire.

The Consumer Product Safety Commission maintains a variety of data on
product recalls and incidents related to consumer products. However, none of
the four databases that it maintains can identify information about federal
facilities or federal employees.

The U. S. Fire Administration is chartered as the nation?s lead federal
agency for coordinating fire data collection and analysis. However, the
national fire incident databases maintained by the U. S. Fire Administration

Page 10 GAO- 01- 879 Fire Safety

do not gather specific information about whether a fire occurred on private
or government property or whether the fire involved specific products. The
Fire Administration maintains the National Fire Incident Reporting System
(NFIRS)- a national database through which local fire departments report
annually on the numbers and types of fires that occur within their
jurisdictions, including the causes of those fires. Reporting, however, is
voluntary; according to the U. S. Fire Administration, this results in about
one- half of all fires that occur each year being reported. In addition, the
U. S. Fire Administration does not collect data on the number of fires in
federal office facilities and the causes of those fires, nor about specific
types of products involved in a fire. According to its comments on a draft
of our report, the Fire Administration does not have the resources or
authority to implement a nationwide study of fires in federal workspace.

In addition to the federal databases, NFPA also maintains a national fire
incident database. According to NFPA, between 1993 and 1997, an average of
6,100 fires occurred per year in federal and nonfederal office space,
resulting in an average of 1 death, 91 injuries, and $131.5 million in
property damage per year. NFPA?s estimates are based on information that
fire departments report to the Fire Administration?s NFIRS system and on
information from NFPA?s annual survey. NFPA annually samples the nation?s
fire departments about their fire experiences during the year; using this
data, NFPA projects overall information about fires and their causes to the
nation as a whole. However, neither the U. S. Fire Administration nor NFPA
gathers specific information about whether a fire occurred on private or
government property or whether the fire involved specific products.

In the past, the federal government has collected data regarding fires
occurring on federal property. The Federal Fire Council was originally
established by Executive Order within GSA in 1936 to act as an advisory
agency to protect federal employees from fire. The council was specifically
authorized to collect data concerning fire losses on government property.
However, the council moved to the Department of Commerce in 1972 and was
abolished in 1982.

Page 11 GAO- 01- 879 Fire Safety

Along with manufacturers, consumer representatives, fire safety officials,
and others, the federal government is one of several important stakeholders
involved in the standards- development process. However, as previously
discussed, the government does not consistently and comprehensively collect
information on fire incidents in federal facilities, and hence it cannot
systematically provide these data to standardsdevelopment organizations for
consideration during revisions of standards. Furthermore, some federal
agencies may be slow to respond to information about failures of certain
products, including those products intended to suppress fires. In at least
one case, a fire sprinkler product that failed in both the work place and
the testing laboratory, as early as 1990, continued to be used in federal
facilities, and it has only recently been replaced at some facilities. This
case is discussed below.

Omega sprinklers were installed in hundreds of thousands of nonfederal
facilities and in about 100 GSA- managed buildings. 3 In 1990, a fire
occurred at a hospital in Miami, FL, resulting in four injuries. During this
fire, Omega sprinklers failed to activate. Through 1998, at least 16
additional fires occurred, during which Omega sprinklers failed to work,
including a May 16, 1995, fire at a Department of Veterans Affairs hospital
in Canandaigua, NY. During the New York fire, an Omega sprinkler head
located directly over the fire failed to activate. Losses resulting from
these and other fires were estimated at over $4.3 million (see table 1).

3 Between 1982 and 1998, Central Sprinkler Company (CSC), one of the largest
suppliers of fire sprinklers, manufactured between 9 to 10 million
sprinklers under the brand name Omega. The Federal Government

Does Not Have Standard Procedures for Collecting Data on Fires in its
Facilities or for Sharing This Information With Standards- Development
Organizations

The Case of Omega Sprinklers

Page 12 GAO- 01- 879 Fire Safety

Table 1: Known Fires in Which Omega Fire Sprinklers Failed to Activate Year
Location Facility Injuries Estimated

loss ($)

July 1990 Miami, FL Hospital Yes, 4 $600 May 1993 Simi Valley, CA Private
home No * Nov 1994 Three Rivers, TX Federal correctional

facility ** Jan 1995 Romulus, MI Hotel Yes $5,000 May 1995 Canandaigua, NY
Veterans Affairs

Medical Center No a $1,000 Jan 1997 West Hollywood, CA Apartment No $15,000
April 1997 Fort Wayne, IN Juvenile holding

facility No $80,000 Sept 1997 Gulf Breezes, FL Marina No $3,500,000 Oct 1997
Kent Island, MD Bed and breakfast No $50,000 Nov 1997 Horsham, PA Hotel No *
Feb 1998 Marrietta, GA Lodge No $3,000 Feb 1998 Milford, MA Rooming house No
$250,000 March 1998 Beverly, MA Dormitory No $5,000 March 1998 Riverside, CA
Private residence * * May 1998 Escondido, CA Retirement center No * May 1998
Scottsdale, AZ Apartment Yes $400,000 * Dallas, TX Private residence * *

Total $4,309,600

*Data were not available. a One injury was sustained by a staff member
during the use of a fire hydrant while helping to

extinguish the fire. Source: GAO analysis based on files reviewed at the
Fairfax County, VA, Fire Department and interviews with County officials, as
well as interviews with officials from, and/ or documents maintained by, the
Marriott Corporation; the Consumer Product Safety Commission; GSA; the
National Institute of Standards and Technology; the Department of Veterans
Affairs; and Underwriters Laboratories.

Although none of the fires reported in table 1 occurred in Fairfax County,
VA, the County fire department became concerned that many of the sprinklers
were installed in public and private facilities in the county. Throughout
the mid- 1990s, by publicizing its concerns about the sprinklers, the County
fire department contributed to the widespread dissemination of information
about the sprinklers in the media. In addition, tests performed in 1996 at
independent testing laboratories- Underwriters Laboratories and Factory
Mutual Research Corporation- revealed failure rates of 30 percent to 40
percent.

On March 3, 1998, the Consumer Product Safety Commission announced that it
had filed an administrative complaint against the manufacturer, resulting in
the October 1998 nationwide recall of more than 8 million

Page 13 GAO- 01- 879 Fire Safety

Omega sprinklers. The agency began investigating Central Sprinkler Company?s
Omega sprinklers in 1996 when an agency fire engineer learned about a fire
at a Marriott hotel in Romulus, MI, where an Omega sprinkler failed to
activate. After identifying that there was a hazard that warranted recalling
the product, the Commission staff sought a voluntary recall from Central.
Unable to reach such an agreement with Central, the agency?s staff were
authorized to file an administrative complaint against the company.
Moreover, the Commission attempted to coordinate with other federal
agencies, such as the Department of Veterans Affairs and GSA. The Department
of Veterans Affairs participated in the recall in accordance with the terms
of the Commission?s settlement agreement with the manufacturer.

GSA officials stated that they became aware of the problems associated with
Omega sprinklers in 1996 after hearing about them from the news media and
Fairfax County Fire Department officials. GSA began a survey to identify the
100 GSA- managed buildings that contained the sprinklers. It also pursued an
agreement with the manufacturer, resulting in a 1997 negotiated settlement
for the replacement of some 27, 000 devices in GSAcontrolled buildings.

Officials from OSHA stated that they were unsure about when they became
aware of the problems associated with Omega sprinklers. An agency official
explained that OSHA generally does not monitor information regarding
problems with specific products, except for Consumer Product Safety
Commission recalls. According to OSHA, it checks such recalls only
informally and within the limited context of one of its programs, but not as
a part of its primary compliance efforts. In addition, according to OSHA
officials, when OSHA did find out about the Omega sprinklers problems, it
took no action because such problems are outside the agency?s jurisdiction
unless the problems involve noncompliance with applicable OSHA requirements.
According to an OSHA official, OSHA does issue ?Hazard Information
Bulletins? that could potentially contain information about failures of
specific products. However, these bulletins do not generally duplicate
Consumer Product Safety Commission recall information and do not generally
concern consumer products.

Federal facilities not controlled by GSA- including those of Capitol Hill
(the House of Representatives, the Capitol, the Senate, and the Library of
Congress) and the Smithsonian Institution- have either recently replaced or
are just now replacing the defective Omega sprinklers. According to an
official of the Architect of the Capitol, although the facility?s management

Page 14 GAO- 01- 879 Fire Safety

was aware of the problems with the sprinklers, it continued using them
because of cost considerations. At the time our review was completed, the
Architect of the Capitol had removed and replaced the Omega sprinklers from
all of the House of Representatives buildings and Capitol buildings, most of
the Senate buildings, and one of the Library of Congress? buildings. The
Architect of the Capitol was also in the process of replacing them in the
remainder of the Senate and Library buildings. In addition, according to the
Chief Fire Protection Engineer of the Smithsonian, agreement for a free- of-
cost replacement of the Omega sprinklers has been reached, although the
process of replacing them had not begun at the time we completed our work.

At your request, we also reviewed concerns about the extent to which
information technology equipment- such as computer printers, monitors, and
processing units- could be a source of fires in offices, homes, and other
places, including federal workspace. A private testing laboratory in Sweden
recently performed experiments that suggested that some types of information
technology equipment could be subject to damage from flames that originate
from external sources. In response to these concerns, the Information
Technology Industry Council convened a panel of stakeholders- including the
Consumer Product Safety Commission, Underwriters Laboratories, and others-
to study the issue. The panel found that information technology equipment
did not pose a widespread fire threat in the United States. According to the
representatives of the American Chemistry Council, the threat of information
technology equipment fires from external sources is mitigated by the
presence of various types of flame retardants in the casings of this
equipment. Moreover, representatives of the Information Technology Industry
Council stated that the industry has a policy of making its equipment as
safe as possible for consumers. They agreed, however, that the issue of the
flammability of information technology equipment needed further study.

Fires, even relatively small ones, can have tragic and costly consequences.
Knowing the numbers and types of fires in workspace, as well as the causes
of fires and any products involved, is critical for understanding the extent
of the risk of fire and can lead to identification and implementation of
steps to reduce this risk. Some private sector organizations- for example, a
major hotel chain and some insurance organizations- track the number of
fires in different types of facilities and their causes. Such information is
used to manage this risk and reduce property damage, injuries, and the loss
of life. However, the federal government, which employs over two million
people in space that GSA and other agencies Conclusion

Page 15 GAO- 01- 879 Fire Safety

manage, collects very limited information on fires and lacks information on
the risk of fires in its workspace. Without more complete information on
fires, the federal government- a key player in the standardsdevelopment
process- cannot provide timely information on the causes of fires in federal
facilities to standards- development organizations for their use in
developing and revising standards, testing procedures, and certification
decisions. Collecting and analyzing data on the risk of fire in its
workspace could enable the government to better protect its employees and
enhance its ability to participate in producing standards that would better
protect the public at large from fire.

We recommend that the Administrator, U. S. Fire Administration, in
conjunction with the Consumer Product Safety Commission, GSA, OSHA, and
other federal agencies that the Fire Administration identifies as being
relevant, examine whether the systematic collection and analysis of data on
fires in federal workspace is warranted. If they determine that data
collection and analysis are warranted, data that should be considered for
collection and analysis include: the number of fires in federal workspace;
property damage, injuries, and deaths resulting from such fires; and the
causes of these fires, including any products involved. In addition, the
agencies should discuss, among other topics deemed relevant, the
availability of resources for implementing any data collection system and
any needed authority to facilitate federal agencies? cooperation in this
effort.

We provided copies of a draft of this report to the heads of the Federal
Emergency Management Agency?s Fire Administration and GSA, as well as the
Consumer Product Safety Commission and the Department of Labor. Because of
its role in testing Omega sprinklers, we also provided a copy of the report
to Underwriters Laboratories. Although Underwriters Laboratories had no
comments on the draft, the other recipients of the draft provided comments
via E- mail. These comments, and our responses to them, are discussed below.

In commenting on our draft report, the Director of the Fire Administration?s
National Fire Data Center agreed in principle with our recommendation by
stating that Fire Administration officials would gladly meet with GSA and
others to examine whether specialized data collection is warranted. We
welcome the Fire Administration's proposal. In addition, the Fire
Administration listed several obstacles to the creation of a complete and
accurate fire incident reporting system: (1) its lack of resources, (2) its
lack of authority to require other federal agencies to Recommendation for

Executive Action Agency Comments and Our Evaluation

Page 16 GAO- 01- 879 Fire Safety

report fires, and (3) its lack of on- site management and control over an
existing fire incident reporting system, the National Fire Incident
Reporting System (NFIRS). Moreover, the Fire Administration does not
specifically collect data on the number and causes of fires in federal
office facilities, and no indication exists that the fire problem in federal
facilities differs significantly from the overall national fire experience
in similar workplace environments. We agree that data on federal fires are
not currently collected, and we would cite this lack of information as a
significant reason for exploring the need for a system to report the number
and causes of fires in federal space. We further agree that a lack of
resources, of authority to compel fire incident reporting, and of management
over reporting may pose serious obstacles to improved fire incident
reporting; therefore, we urge that the Fire Administration address these
factors with other agencies when it meets with them to discuss the need for
more specialized reporting on fires in federal work space.

GSA senior program officials commented on a draft or our report. They
requested that we delete a statement in our draft report that GSA could not
provide us with complete information on fires that caused over $100,000
damage in federal facilities it manages. GSA said that our statement was not
germane. We declined to make this change because the statement is germane to
our discussion about a lack of information on fires in the federal
workplace. GSA?s inability to provide the information we requested serves to
illustrate this very point. In addition, we added information in our report
regarding GSA?s explanation that it had cancelled a previous requirement for
its regional offices to report smaller fires to a central repository. GSA
explained that such reporting was labor intensive and time consuming, and
analyses of this information could not yield specific fire trends. We agree
with GSA that some reporting requirements may be labor intensive, time
consuming, and not helpful. Therefore, in our view, as stated above and as
reflected in our recommendation, the Fire Administration should address
these factors with GSA and other agencies when it meets with them to discuss
the need for more specialized reporting on fires in federal work space. GSA
did not comment on the recommendation in the draft of our report.

In addition, Department of Labor officials provided technical and clarifying
comments, all of which we incorporated into our report. However, they did
not comment on the recommendation. The Department of Labor?s Bureau of Labor
Statistics Assistant Commissioner, Office of Safety and Health, provided
additional data regarding the number of federal employees who died as a
result of fires or explosions from 1992 through 1999, clarifying that most
of these fatalities occurred outside of

Page 17 GAO- 01- 879 Fire Safety

federal buildings. The Department?s Occupational Safety and Health
Administration?s Acting Director for Policy provided additional information,
which we incorporated into our report, about the extent of its involvement
in the Omega sprinkler case and the rationale for the actions it took. The
Consumer Product Safety Commission stated that its comments were editorial
in nature, and we revised our report to incorporate these comments.

As arranged with your office, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days after
the date of this letter. At that time, we will send copies of this report to
the cognizant congressional committees; the Administrator, General Services
Administration; the Chairman, Consumer Product Safety Commission; the
Secretary of Labor; and the Administrator, Federal Emergency Management
Agency. We will also make copies available to others on request.

If you have any questions about this report, please contact me at (202) 512-
4907. Key contributors to this report were Geraldine Beard, Ernie Hazera,
Bonnie Pignatiello Leer, Bert Japikse, and John Rose.

Sincerely yours, Peter Guerrero Director, Physical Infrastructure Issues

Appendix I: Objectives, Scope, and Methodology

Page 18 GAO- 01- 879 Fire Safety

Our report (1) provides information on the federal government?s reliance on
private voluntary fire standards and testing products against those
standards and (2) discusses whether data that are available about fire
incidents and their causes in civilian federal facilities are sufficient to
protect federal workers from the threat of fire.

To examine the government?s reliance on fire safety standards and testing,
we reviewed policies and procedures regarding how standards- setting
organizations and independent laboratories establish fire safety standards
and test products, as well as the roles of federal agencies and other
interested parties in these processes. We contacted standardsdevelopment
organizations, including Factory Mutual Research, Underwriters Laboratories,
Southwest Research Institute, the American National Standards Institute
(ANSI), and the American Society for Testing and Materials. We also obtained
information regarding how testing and standards- setting laboratories and
organizations consider fire incident data and other information about fire
hazards when revising fire safety standards and testing procedures. We
obtained and analyzed regulatory and statutory criteria regarding the
federal role in fire safety standards and testing. We interviewed federal
officials from the General Services Administration (GSA), the National
Institute of Standards and Technology, the U. S. Fire Administration, the
Consumer Product Safety Commission, and the Department of Labor, as well as
officials from standardsdevelopment organizations. We also interviewed fire
protection officials, including officials from the International Association
of Fire Fighters, the International Association of Fire Chiefs, and the
Fairfax County, VA, Fire Department to obtain information on setting
standards and testing products.

To examine whether data are available about incidents and causes of fires in
civilian federal facilities, we contacted GSA, the manager of about 40
percent of all civilian, federal office space. However, GSA does not
routinely collect information about all fires that occur in federal
facilities. Therefore, we obtained and analyzed fire protection incident
data from the Fire Administration and the National Fire Protection
Association (NFPA). The U. S. Fire Administration maintains the National
Fire Incident Reporting System, which is the world?s largest national annual
database of fire incident information. State participation is voluntary,
with 42 states and the District of Columbia providing reports. The data in
the National Fire Incident Reporting System comprise roughly one half of all
reported fires that occur annually. NFPA annually surveys a sample (about
onethird) of all U. S. fire departments to determine their fire experiences
during the year. NFPA uses this annual survey together with the National
Appendix I: Objectives, Scope, and

Methodology

Appendix I: Objectives, Scope, and Methodology

Page 19 GAO- 01- 879 Fire Safety

Fire Incident Reporting System to produce national estimates of the specific
characteristics of fires nationwide. Through a review of the databases, we
found that there was not sufficient detail to determine which of the fires
reported occurred in federal facilities. In addition, the fire departments
do not document the name brands of any product that might have been involved
in a fire. However, NFPA was able to provide information about fires that
have occurred in office space (federal and nonfederal) from 1993 through
1998. Finally, we did not conduct a reliability assessment of NFPA?s
database or the National Fire Incident Reporting System.

We also attempted to determine the number of civilian federal employees who
may have been injured or killed as a result of a fire- related incident
while at work. In this regard, we obtained information from the Bureau of
Labor Statistics? Census of Fatal Occupational Injuries (CFOI) regarding
civilian federal employee fatalities from 1992 through 1999. The federal
government and the states work together to collect work- related fatality
data from workers? compensation reports, death certificates, news stories,
and other sources for CFOI. All 50 states participate in CFOI. The Bureau of
Labor Statistics was able to provide information from CFOI describing the
number of civilian federal employees fatally injured due to fire- related
incidents while at work. We also obtained information from the Office of
Workers? Compensation Programs from 1992 through April 2001 regarding
civilian federal employees or their families who have filed for workmen?s
compensation as a result of an injury or fatality due to a fire- related
incident while at work. However, the data represent only those incidents for
which a civilian federal employee or the family filed a claim. With the
limited data available from the fatal injuries census and Office of Workers?
Compensation Programs, we were unable to do an analysis of the number of
claims filed due to bombings, such as the April 1995 Murrah Federal Building
bombing in Oklahoma City, OK, and the August 1998 bombing of the U. S.
Embassy in Dar Es Salaam, Tanzania. In addition, according to CFOI, the
fatality data do not include fatalities due to bombings, such as the
Oklahoma City bombing and the Dar Es Salaam bombing. When a fatality is
reported, CFOI requires that Assaults and Violent Acts, Transportation
Accidents, Fires, and Explosions reports take precedence in the reporting
process. When two or more of these events occur, whoever inputs the
information selects the first event listed. The Bureau of Labor Statistics
classified the Oklahoma City bombing deaths as homicides under the Assaults
and Violent Acts category. In addition, the Office of Workers? Compensation
Programs was able to provide information on the number of injuries to
civilian federal employees that its Dallas District Office reported for 1995
as resulting from explosions.

Appendix I: Objectives, Scope, and Methodology

Page 20 GAO- 01- 879 Fire Safety

According to the Office of Workers? Compensation Programs, it is likely that
many of these injuries resulted from the Oklahoma City bombing. Furthermore,
the databases do not contain any details of fires. We used the fatality data
from CFOI, because it is the more comprehensive source of federal employee
fatality information. Finally, we did not conduct a reliability assessment
of the Bureau of Labor Statistics? CFOI database or the database of the
Office of Workers? Compensation Programs.

We also obtained information about fire incidents related to consumer
products by contacting the Consumer Product Safety Commission. The
Commission maintains several databases that allow it to conduct trend
analyses of incidents involving various types of products, including the
National Electronic Injury Surveillance System, a Death Certificate File,
the Injury or Potential Injury Database, and the In- Depth Investigation
File. In addition, the Commission maintains a library (paper files) of
information on products that have been recalled. However, none of these
sources contained information that would identify information about federal
facilities, federal employees, or product brand names, with the exception of
those that have been recalled. To examine the quality and limitations of
these data, we reviewed relevant documents and interviewed officials from
organizations that compile and report the data, including the National Fire
Protection Association, Fire Administration, Consumer Product Safety
Commission, Occupational Safety and Health Administration, Bureau of Labor
Statistics, Office of Workers? Compensation Programs, and National Institute
of Standards and Technology.

As requested, we examined details about reporting incidents and concerns
involving Omega sprinkler heads and how standards- development
organizations, federal agencies, and others responded to reports about the
failures of these devices. We contacted officials from, and in some cases
obtained documentation from, the Fairfax County (VA) Fire Department. We
also contacted various federal regulatory agencies or agencies that used or
were indirectly involved in using Omega sprinklers, including GSA, the
Consumer Product Safety Commission, Occupational Safety and Health
Administration, National Institute of Standards and Technology, Architect of
the Capitol, Smithsonian Institution, and Department of Veterans Affairs. We
also contacted officials from various laboratories that had tested Omega
sprinklers, including Underwriters Laboratories, Factory Mutual, and the
Southwest Research Institute. We also interviewed officials from the
Marriott Corporation, which, along with Fairfax County, had publicized the
problems associated with the sprinklers.

Appendix I: Objectives, Scope, and Methodology

Page 21 GAO- 01- 879 Fire Safety

As requested, we also reviewed concerns about the possible flammability of
information technology equipment. In this regard, we inquired and obtained
information about such factors as the types of flame retardants currently
used in the casings of information technology equipment and concerns about
the environmental and health impacts of these substances, the standards used
to mitigate the flammability of information technology equipment, and the
tests used to determine the flammability of this equipment. Our sources of
information were the American Chemistry Council; the Great Lakes Chemistry
Council; the Information Technology Industry Council; the National
Association of State Fire Marshals; SP (a private testing laboratory in
Sweden); the National Fire Protection Association; Underwriters
Laboratories; and federal agencies, including the U. S. Consumer Product
Safety Commission and the U. S. Department of Commerce?s National Institute
of Standards and Technology.

We conducted our work from December 2000 through August 2001 in accordance
with generally accepted government auditing standards.

Appendix II: Sources Contacted by GAO Page 22 GAO- 01- 879 Fire Safety

Architect of the Capitol Consumer Product Safety Commission Department of
Labor:

Bureau of Labor Statistics Occupational Safety and Health Administration
Office of Workers? Compensation Programs Department of Veterans Affairs
Federal Emergency Management Administration:

United States Fire Administration General Services Administration:

Federal Supply Service Inspector General Public Buildings Service Library of
Congress National Institute of Standards and Technology Smithsonian
Institution

American National Standards Institute American Society for Testing and
Materials Factory Mutual Research National Fire Protection Association
Southwest Research Institute Appendix II: Sources Contacted by GAO

Federal Agencies StandardsDevelopment Organizations and Independent
Laboratories

Appendix II: Sources Contacted by GAO Page 23 GAO- 01- 879 Fire Safety

Underwriters Laboratories, Inc. SP Swedish National Testing and Research
Institute

American Chemistry Council Great Lakes Chemistry Council Information
Technology Industry Council Insurance Services Office International
Association of Fire Chiefs International Association of Fire Fighters
National Association of State Fire Marshals National Safety Council Society
of Fire Protection Engineers

Fairfax County Fire Department, Fairfax County, VA Marriott Corporation
Industry Groups

Other Organizations

(395002)

The first copy of each GAO report is free. Additional copies of reports are
$2 each. A check or money order should be made out to the Superintendent of
Documents. VISA and MasterCard credit cards are also accepted.

Orders for 100 or more copies to be mailed to a single address are
discounted 25 percent.

Orders by mail:

U. S. General Accounting Office P. O. Box 37050 Washington, DC 20013

Orders by visiting:

Room 1100 700 4 th St., NW (corner of 4 th and G Sts. NW) Washington, DC
20013

Orders by phone:

(202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone phone. A recorded menu
will provide information on how to obtain these lists.

Orders by Internet

For information on how to access GAO reports on the Internet, send an email
message with ?info? in the body to:

Info@ www. gao. gov or visit GAO?s World Wide Web home page at: http:// www.
gao. gov

Contact one:

 Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm  E- mail:
fraudnet@ gao. gov  1- 800- 424- 5454 (automated answering system) Ordering
Information

To Report Fraud, Waste, and Abuse in Federal Programs
*** End of document. ***