Springfield Daughters of the American Revolution. Illinois. Sp.

Statistical studies on fire incidents involving textiles online

. (page 1 of 2)
Online LibrarySpringfield Daughters of the American Revolution. Illinois. SpStatistical studies on fire incidents involving textiles → online text (page 1 of 2)
Font size
QR-code for this ebook


94 34



STATISTICAL STUDIES ON
FIRE INCIDENTS INVOLVING
TEXTILES



^'\/^-n o-^o (4ia^



STATISTICAL STUDIES
ON
FIRE INCIDENTS INVOLVING TEXTILES



HENRY JDVEY



NATIONAL FIRE PREVENTION AND CONTROL
ADMINISTRATION

U.S. DEPARTMENT OF COMMERCE



PROPERTY OF WKk .

NATIONAL FIRE REFERENCE SERVICE
U.S. Dept. of Commerce
Washington, D. C. 20230



Prepared for presentation at the

14TH INTERNATIONALE CHEMIFASERTAGUNG 1975

September 24-6, 1975, Dornbirn, Austria



,*^,f^>'^^ . ^'1^''



STATISTICAL STUDIES

ON

FIRE INCIDENTS INVOLVING TEXTILES



1,' Introduction

I am very happy to be here today and proud to have
been invited to give this presentation. It is both a
pleasure and an honor to be on the program of this Conference.

Because it is a special honor to be invited to speak
before you, I have given much thought to this presentation.
I was told that the audience will be composed of high-level
representatives of the fiber, textile, and garment industries
with a sprinkling of high government officials and executives.
Such an audience, it seemed to me, would not be interested in
mere numbers - in learning how many incidents of a certain
type we have in the USA. But yet, the subject of my presenta-
tion deals with textile fire statistics, and it obviously
must deal with numbers. What I have decided to do, therefore,
is to emphasize the methodological aspects' of my topic rather
than the findings. Accordingly, I will start with a description
of the Flammable Fabrics Accident Case and Testing System.

FFACTS , as we call the system, has been the major single
source of data on fabric fire incidents during the past five

years, and it is important that you know something about its

strengths and limitations. I will then review for you the



-2-

analytical methodology used in the development of priorities
for flammability standards. Next I will discuss the two
major statistical systems dealing with data on textile fires
in the USA. And finally, I will bring you up to date on
recent developments on the international level.



-3-

2. The Flammable Fabric Accident Case and Testing System .

The Flammable Fabrics Act ^Ij authorized the promulgation
of flammability standards for fabric products to protect the
public against unreasonable risk of fire leading to death,
injury, or significant property damage. In 1973 the responsi-
bility for this Act was assigned to the Consumer Product
Safety Commission \2\ . The National Bureau of Standards
provides technical support to the Commission, and up to 1975
the Flammable Fabrics Accident Case and Testing System was
included in this support. At that time it was transferred to
the Commission.

As initially conceived, FFACTS was to provide statistical
data on deaths, injuries, and economic losses resulting from
"unreasonable risks", as required by law, and to aid in
establishment of priorities for standards JsJ . It was also
planned that it would develop information on the behavior of
fabric products in fire incidents , to help in designing small-
scale tests that could predict real-life behavior; and that
it would contribute to the' common store of/ knowledge on the
nature of fabric flammability. As it turned out, it was also
instrumental in developing information on the role played by
ignition sources in flammable fabric incidents, and it aided
in developing suggestions on how they can be designed for
improved safety.



-4-

Input for FFACTS are in-depth investigative reports
on flammable fabric incidents. Using specially designed
forms , investigators around the country follow and report
on accidental fires involving textile products . These
repKjrts are screened to be sure that they are in fact
relevant and that they contain enough information to make
processing worthwhile. If found acceptable, the reports are
carefully analyzed and the information derived from them,
together with data from laboratory tests performed on any
remains of the involved fabric products received, are encoded
and computerized.

Because the investigations do not follow a statistically
designed sampling plan, the FFACTS data base is not a statis-
tically representative sample of all fabric fires in the
United States. However, many trends which began to develop
in the early data have remained consistent while the geograph-
ical sources and the collection methods have changed several
times, and the data base doubled and tripled in size. It
thus appears that FFACTS data reflect the actual fire experience
in the United States.



3. Selection of Priorities for "Standards .

By the time the FFACTS data base contained two thousand
incident reports, the methodology applied to analyze the data
had been developed into a formal protocol [4^ . The process
of developing priorities for flammable fabrics standards
called for successive examination of the data in the light
of nine criteria, with each step producing a more refined
stratification and eventual identification and ranking of
the most severe hazards.

TABLE 1. [4i

CRITERIA APPLIED IN DEVELOPING
PRIORITIES FOR FLAMMABLE FABRIC STANDARDS

* Definition of population

* Flammable liquid involvement

* Distribution by item category

* Distribution by sex

* First-to-ignite

* Severity of injury

* Age group

* Technological practicability

* Regulatory feasibility

The first step called for defining the population to
be studied. In the example I will be using, that population
consisted of the 1,492 garment-involving incidents contained



-6-

in the December 1972 FFACTS data. base of 1,964 incidents
reports.

The second decision concerned incidents in which contam-
ination with a flammable liquid, such as gasoline, may have
obscured the role played by the garment. Because such
incidents do not help in the identification of intrinsically
hazardous garments , they were eliminated from further
consideration .

The remaining incidents were then distributed by garment
category. Sleepwear ranked first in frequency, with underwear,
shirts, pants, and dresses following in that order.

TABLE 2. [_4]]

Most Frequently Involved Garment Categories
(FFACTS, December 1972)

Garment Number of Incidents



Sleepwear 391

Underwear 349

Shirts/Blouses , 255

Pants 152

Dresses 120



.7-

Distributing the data by sex and frequency put women's
sleepwear in the first place and dropped men's sleepwear to
the bottom of the list. Underwear was still high on the
list, with men's underwear in the second, and women's in the
fourth place.

TABLE 3. ^43

Most Frequently Involved Garment Categories
by Sex
(FFACTS, December 1972)

Garment Sex Total Incidents



Sleepwear


Female


298


Underwear


Male


212


Shirts


Male


200


Underwear


Female


137


Pants


Male


129


Dresses


Female


119


Sleepwear


Male


93



The application of the next criterion in the selection
process resulted in the elimination of all garments which were
not first-to-ignite from further consideration. This criterion
was applied on the assumption that the first garment ignited
largely determines the hazard. It should be pointed out,
however, that while this assumption may be reasonable, it
has not been proven. Moreover, because the order of ignition
is often not determinable in the more severe cases, particularly



-8-

if the victim does not live to tell the story, the application
of this criterion tends to introduce a bias.

The elimination of all garments which were not first-to-
ignite produced a major change in ranking. Underwear, which
was often involved but rarely as the first ignited item,
dropped to* the bottom of the list.

TABLE 4. [4]

Most Frequently Involved First-to-Ignite
Garment Categories by Garment type/ Sex group
(FFACTS, December 1972)







Total


Number First


Garment


Sex


Incidents


to Ignite


Sleepwear


Female


298


243


Shirts


Male


200


120


Dresses


Female


119


78


Sleepwear


Male


93


75


Pants


Male


129


42


Underwear


Male


212


23


Underwear


Female


137


4



The next criterion in the selection process deals with
injury severity. However, while this criterion was considered
in 1972, it was not actually applied, because we did not know
how. It is now 1975, and though much work has been done on
this problem, the relationship between injury severity and
fabric or garment parameters is still not fully elucidated.

A study on the February 1974 FFACTS data base of 3,047
incident reports, showed that the cover and fit of a garment
correlate with extent of burn injury significantly, with shirts



-9-

and pants showing least extensive burn injuries, pajamas
intermediate, and dresses, nightgowns, and robes showing
most extensive burns fsj . Of the other fabric/garment
parameters studied, only fiber content showed a significant
correlation, with garments made from all-cellulosic and
cellulosic-blend fabrics showing a correlation with more
extensive burns than those made from single-layer fabrics
made from thermoplastic fiber. Multi-layered garments from
thermoplastic fiber fabrics tended to behave more like the
cellulosics than like single-layer thermoplastics. Fabric
weight, construction, surface characteristics, and burn time
showed no consistent correlation with extent of iniury. It
must be remembered, however, that many of the large-cover
garments, such as robes and dresses, tend to be worn by
older people and by females, and their defensive capabilities
are in general not as good as those of younger people and
males. For this reason the question of the intrinsic hazard
of such garments is still not fully answered. Accordingly,
we are still unable to apply an injury severity criterion
to the priority development process.

Age of the victim was the next criterion applied, and
to make the data more meaningful the age continuum was divided
into age groups believed to exhibit broadly similar behavior
patterns. Of course, the fractions of the total population
in each of these sex-age groups are not equal and the data



-10-

had to be rationalized to permit rank ordering. This was
done by dividing the number of garments in a given sex-age
group by the proportion of the U.S. population represented
by that sex-age group. Thus, for example, the first- to-
ignite sleepwear/ female incidents were distributed by age
group. This showed that there were 50 incidents in the 0-5
age group. This number was divided by .051, since girls
in the 0-5 age group constitute 5.1% of the total US population.
The resulting number, 98, is the highest on the list and
indicates the highest hazard level.

The values obtained in this manner are called relative
incident frequency indexes because they show the relative
frequency of involvement of various garment types for different
portions of the population. They can be compared across age
groups and garment types, and thus make possible the achieve-
ment of the objective of the entire process, the rank-ordering
of candidates for garment flammability standards in accordance
with the hazard they represent.

In the example we hav& been following, this ranking
put sleepwear in the first four places - sleepwear for females
in the 0-5 group first, with an index of 98, sleepwear for
females in the 6-12 age group second, with an index of 87,
sleepwear for females 66 years old and older third, and sleep-
wear for males in the 0-5 age group fourth. The next highest
category, dresses for girls in the 0-5 age group have an index
of 49, one half that of the first-ranking item.



-11-



TABLE 5



a



RELATIVE INCIDENT FREQUENCY INDEX VALUES
FOR GARMENT TYPE -SEX- AGE GROUPS



Garment


Sex


Age


Index






Group


Value


Sleepwear


Female


0-5


98


Sleepwear


Female


6-12


87


Sleepwear


Female


66+


71


Sleepwear


Male


0-5


69


Dresses


Female


0-5


49


Sleepwear


Female


46-65


40


Shirts


. Male


66+


35


Dresses


Female


0-12,66+


29 (ea)


Shirts


Male


13-20


26


Sleepwear


Female


21-45


24


Shirts


Male


21-45


24


Shirts


Male


46-65


23


Sleepwear


Male


66+


23


Sleepwear


Male


6-12


22


Shirts


Male


0-5


21


Shirts


Male


6-lZ


18


Sleepwear


Female


13-20


14


Pants


Male


0-5,66+


13 (ea)


Pants


Male


46-65


11


Dresses


Female


46-65


9


Pants


Male


6-12


8


Pants


Male


21-45


7


Sleepwear


Male


46-65


7


Sleepwear


Male


13-20


5


Pants


Male


13-20


4


Dresses


Female


21-45


4


Sleepwear


Male


21-45 '


43


Dresses


Female


13-20


1



The groups ranked high by the relative incident frequency
index values represent the highest hazards and constitute
priorities for flammability standards. However, standards
must be technologically practicable, and the ability of the
industry to provide the necessary flame resistant garments
must be given due consideration. Of course, three out of the



-12-

first four highest ranking groups identified in this analysis
are already covered by standards. The fourth one, sleepwear
for females 66 years and older, presents a regulatory feasi-
bility problem, since sleepwear for the elderly cannot be
distinguished from other adult sleepwear, and there seems
to be no way for a mandatory regulation to include that
segment of the population, and exclude the remainder which
need not be made subject to the regulation.

It would be possible to apply a regulation across the
board, as for example to all sleepwear for females. However,
the desirability of such action has been questioned, even
if it were considered technologically practicable. According
to preliminary study by the Bureau of Economic Analysis of the
Consumer Product Safety Commission, treating the fabric
yardages used in nightwear, shirts, and trousers so that
they can meet a vertical self -extinguishing test similar to
FF5-74 fs} would increase the retail value of the apparel
by from 2.5 to 5 billion dollars i7j . The direct cost of
treating sufficient yardage to replace the garments that
would fail such a test were estimated to be $310 million
for women's nightgowns, $81 million for robes and housecoats,
$72 million for men's pajamas, $340 million for dresses,
$786 million for men's and boys' shirts, and $516 million
for men's and boys' trousers. These numbers underscore the
need for an accurate determination of the magnitude of textile



-13-

flainmability hazards in terms which permit comparison
with other hazards and with costs of remedial actions. This
FFACTS cannot do, because the priorities that can be derived
from an analysis of FFACTS data are relative priorities , and
they constitute a ranking within the fabric products family.



-14-

4. Statistical Systems . ■-

For this reason, much current effort in the United States
is aimed at supplementing the detailed studies of flammable
fabric incidents, such as those in FFACTS , with systems
ca'pable of providing statistically valid data and placing
the entire textile fire problem on a more quantitative basis.
Of these there are two major ones, the National Electronics
Injury Surveillance System (NEISS) fsj , and the National
Fire Data System (NFDS) ^s] .

NEISS is run by the Consxamer Product Safety Commission.
In essence, it is an injury monitoring system designed to
determine the nature and scope of consumer safety problems.
It is set up as a network of 119 statistically selected
hospital emergency rooms. The data collection begins when
the patient shows up in the emergency room for treatment.
The admission clerk obtains basic information about the
accident, and records it. This information is then coded
and transmitted electronically to the^ CPSC in Washington,
where it is collected and organized.

NEISS data are published monthly in NEISS News. According

to the July 1975 issue, 34 clothing -involving accidents
were reported in April to the emergency rooms of the 119
NEISS hospitals Jj-Oj . Just to give a flavor for the variety
of products listed, I will mention that this compares with
2,239 accidents involving stairs, 1,458 involving bicycles.



-15-

95 involving cooking ranges, and 5 involving water heaters.
Generally, accidents involving fabric products do not rank
high on the NEISS list.

TABLE 6. *

Injuries Treated at NEISS Hospitals -
Association with Selected Consumer Products





Number


By Sex


Product Description


of


cases


%




April


7/14/74








4/30/75


M F


Stairs & Landings


2,239


18,791


42 58


Bicycles


1,458


13,825


65 35


Cooking ranges,








ovens


95


987


45 55


Clothing, day








and night


34


227


41 59


Water heaters


5


93


40 60



^Values for products selected for illustrative
purposes from the July 1975 issue of NEISS
News (Vol. 4, No. 1)

NEISS is well suited to perform the product injury
surveillance function it is meant to perform, but it is not
ideal from the point of view of the textile fire problem.
It covers only those incidents that result in an injury that
is treated in a hospital emergency room. This means that
those who die, those who go directly to burn centers, those
who go to private physicians , and those who don't seek medical
help at all are not visible to the system. Moreover, because
the system was designed to relate only one product to each
incident, the data may be difficult to interpret in terms of



-16-

fire incidents, which must be characterized by two products,
an ignition source and the material ignited.

NEISS is a useful tool, and efforts are now being made
to make it even more useful in the fire area. The CPSC is
rxonning surveys to determine the number of injuries that are
not treated in hospital emergency rooms, and this will permit
them to develop correction factors f or . their estimates. They
have also started to flag fire -involving accidents with a
special code, to permit easy identification.

The National Fire Data System is a vital part of the
new National Fire Prevention and Control Administration's Fire
Data Center \l_l] . When the Congress established the Administra-
tion, in 1974, it set for it several major objectives. Among
them are to :

•k Provide an accurate nationwide analysis
of the fire problem

* Assist in setting priorities

* Identify major problem areas
Vf Determine possible solutions

* Monitor fire losses

The NFDS was established to address these objectives.
To do that, it must quantify the fire problem and its
components, to determine their magnitude. Necessarily, this
requires the determination of the magnitude of the textile
fire problem.

A major subsystem of the NFDS collects fire incident
reports prepared by the fire service. This subsystem was



-17-

tested last year in a pilot-phase operation, and in the
process collected some data. The statistical significance
of that data is not determinable, because of the pilot-phase
nature of the operation, and not much credence should be given
to the actual numbers reported. However, I will cite them
here as example of the kind of data that will come from the
system in the future .

The fire incident reports collected during the pilot
phase effort included questions about causal factors. Among
them was a question concerning type of material first ignited,
and 46,837 reports carried an answer to that question: textiles
pr fabrics were the first materials ignited in 5,552 cases, or
in about 12% of the total. Fabrics and textiles constituted
the fourth largest class of the nine major classes of "type of
material first ignited", ranking only behind flammable or
combustible liquids, natural products, and wood and paper.
Within the fabrics and textiles class, cotton and rayon
textiles accounted for 637o of the total; synthetic fibers
or fabrics were second with 25% . Expressing the same values
in another way, cotton and rayon products were first materials
to ignite in 7% of all fires, and synthetics in 3%.

Another question asked in the pilot-phase fire incident
report concerned the form of material first ignited. From a
data base of 46,686 incidents for which this question was answered,
soft goods and apparel accounted for 3,279 cases, or 7% of the
total.



18



^Z i-hO


0^ ►tJ p.


I-'





M t-i


03


00 JO CO


CO


w W rt


CO


^




ro




1 H- CO


H-


m




3


•T33


o


O


H- H-


i-h


H- HTT




CL O (t)


ft


n>


rt o.^


3




T3


rt


hd


n>


CO


3*






P


o




CO


t-ti




fl)


3
rt




►^



Ln


2


rt


I-h n


^'


►Tl ro


H-


03


UlM ><


^


rt


Ui pcJ rt


CO


fD


Nscn H-


aOT






h -




o


H'


1


ft)


H-


^1


13




CO OP


H-


o


nj


-


3


(D


p-


H-




H-


CO


O'l-' Mjrt




(D


o


c


m


O


3


rt


•i


CL Hi


rt




CO






CO


►la




H-


rt




:rn


d "^




p:


I-'




^




CO


03


rt


(D




n>


CO


3-








00


fC
I-h

03


O
I-h

3








cr 03








i-i


rt








H-


m








O


l-t








CO


f'








-


03




•Tl



-19-



FIGUKE 3



■^^■>^' .*->\v:




Classes of form of material
first ignited
NFIRS - Pilot Phase
46,686 incidents



-20-



FIGUEE 4




Clothing
(on person)

1%



Categories of form of
material first ignited in
the soft goods, apparel class
NFIRS - Pilot Phase
3,279 incidents



-21-

The class of soft goods and apparel includes nine
categories: mattresses and pillows; bedding; linen; clothing
(not on the person) ; clothing (on the person) ; curtains and
drapes; soft goods - not made up; luggage; and other or
undetermined soft goods. The category most frequently noted
was mattresses and pillows. They comprised 40% of the 3,279
incidents in the class. The second highest category was
bedding, which contained 23% of the total for the class.
These two categories account for 5% of all the forms of
materials ignited, a number pointing out the hazard of smoking
in bed. The clothing (not on person) category comprised
177o of the soft goods class, but the clothing (on the person)
category only 1% .

The relatively small number of clothing (on the person)
ignitions points out a shortcoming of the fire incident
subsystem as seen from the textile point of view. Because it
is limited to fires attended by the fire service, it misses
all the garment fires that are put out by the victim or
his family or friends . HoW many of these fires are garment
fires is open to question, but the over-all percentages derived
from the pilot-phase operation do not differ markedly from
the percentages produced by another tool of the NFDS , National
Household Fire Survey ri2j ,

The National Household Fire Survey was conducted in
April 1974 to obtain an overview of the national household fire



-22-

problem. Thirty-three thousand households - a statistically
selected sample of U.S. households - were interviewed and
asked whether they experienced a fire during the previous
12 months. If the answer was yes, they were asked a number
of questions concerning the fire.

The statistical projection for the yearly total of
household fires based on answers to these questions is
5,575,000. Of these, interior furnishings (carpets, rugs,
^curtains , upholstered furniture) were indicated as first-to-
ignite items in an estimated 285,000 incidents, clothing in
159,000, bedding in 115,000, and other fabric items in 69,000
incidents. These add up to a total of 628,000, and that means that
according to the survey textile products constitute about 117o
of all the first-to-ignite items.



TABLE 7*








TOTAL OTJMBER OF HOUSEHOLD EIRES


: 5,575,
628,


,000

.000


100%


ALL TEXTILE FIRES


11%


Interior Fumishirigs


285,


,000


5%


Clothing


159,


,000


3%


Bedding


115,


.000


2%


Other Fabrics


69,


.000


1%


*National Household Fire


Survey







-23-

In terms of dollars, the total loss from fires in which
textile products were first to ignite constituted an even
larger percentage of the projected national total of about $1.5
billion. The loss for clothing was a projected $77 million,
for walls and floor coverings - $283 million, and for
furnishings - $84 million. This adds to a total of $445
million worth of damage from fires where a textile product
was first to ignite, and it constitutes about 28% of the
projected national total.'


1

Online LibrarySpringfield Daughters of the American Revolution. Illinois. SpStatistical studies on fire incidents involving textiles → online text (page 1 of 2)