A Guide to the Upholstered Furniture Fire Problem and TB 117

Below is a featured article from Fire Safety Science News #36 by John R. Hall, Jr. of the National Fire Protection Association

Those of us in fire safety science try to ‘solve’ big parts of the fire problem, where ‘solve’ means a large reduction. Typically, a proposal for a new solution must overcome objections along the following lines:

  • The solution won’t work.  It is unreliable or ineffective.
  • The solution costs too much.  The benefits of are smaller than the costs.
  • Other solutions are better.  Changing behavior or changing some other involved item (e.g., switching between heat source and item ignited) would be better.
  • The solution creates bigger problems (other types of harm) than it removes.
  • The targeted fire problem is not that big and/or is declining rapidly.  It does not need a solution.

Sometimes objections like these are accurate and deserve to be factored into the societal decisions about a particular solution.  However, accurate or not, they will always be raised by parties who prioritize other factors over fire losses.

We have been working on the upholstered furniture fire problem for half a century.  Along the way, we have built a consensus on what solutions work, how well they work, how much they cost, how much they are needed, and what non-fire consequences they have.  More recently, that consensus has come unstuck because of accumulating evidence that some of our “solutions” create serious non-fire problems.

The signal event marking this changing consensus was the 2013 decision in California to delete the small open flame portion of the requirements contained in Technical Bulletin 117.  This is – or was – the only regulation in North America that addressed non-smoldering fires involving upholstered furniture.  Because California is such a large part of the North American market, TB 117 had an impact far beyond the state of California. I will leave it to others to describe the growing concerns with flame retardant treatments of upholstered furniture and associated effects on people’s health.  In the limited space available here, I want to frame the discussion of what we should do next in terms of the size and characteristics of the fire problem.  Our challenge is to develop a best estimate of the challenge we face, and then consider the pros and cons of alternative strategies to address that challenge.

From 1980-1984 to 2006-2010, estimated annual average home fires and losses involving upholstered furniture as the first item ignited declined substantially (see figure).

Figure: Number of US home fires and deaths with upholstered furniture as the first item ignited.

During this decline, civilian deaths declined by 61%, from 1,220 (25% of total home fire deaths) to 480 (19%), and fires declined by 77%, from 29,400 (4% of the total) to 6,700 (2%). Moreover, civilian injuries declined by 68%, from 2,630 (13% of the total) to 840 (7%), and direct property damage, after adjustment for inflation to 2010 dollars, declined by 17%, from $522 million (7% of the total) to $434 million (6%). On the one hand, this is a story of great progress, with hundreds of lives saved each year.  On the other hand, this remains one of the largest parts of the U.S. fire death problem.

U.S. fire statistics are based on the National Fire Incident Reporting System (NFIRS), dating back to 1980.  Starting in about 2003, changes to NFIRS made it possible to estimate fires associated with a product in the role of the most important secondary item ignited.  Using this new data (see 2nd row in the table below), NFPA constructed the following analysis considering all of the large pieces of data in the furniture fire problem. 



Civilian Deaths

Civilian Injuries

Direct Damage (in US$ millions)

 Lighted tobacco product

 1,900 (21%)


 270 (45%)

 320 (29%)

 $97 (17%)

 Open flame from other fire  (secondary item)

 2,200 (25%)

 130 (21%)

 280 (25%)

 $138 (24%)

 Operating equipment

 1,500 (17%)

 70 (12%)

 140 (13%)

 $81 (14%)

 Small open flame

 1,400 (16%)

 60 (10%)

 220 (20%)

 $69 (12%)

 Ember, ash or other or  unclassified hot or smoldering  object

 1,300 (15%)

 60 (10%)

 130 (11%)

 $150 (27%)

 Unclassified, other or multiple  heat source

 600 (7%)

 20 (3%)

 30 (3%)

 $31 (5%)


 8,900 (100%)

 610 (100%)

 1,120 (100%)

 $566 (100%)

Table: Upholstered furniture home fire problem, 2006-2010 averages, by major scenario.

The figure below shows upholstered furniture fire deaths by type of ignition source. This comprehensive overview tells us some useful things. Lighted tobacco products (principally cigarettes, cigars and pipes, but not including matches and lighters) account for 45% of upholstered furniture home fire deaths, dwarfing any other scenario but not dwarfing all other scenarios combined.


The 12% share for operating equipment and the 10% share for ember, ash or other or unclassified hot or smoldering object, both could be treated as likely smoldering ignitions, addressable by a smoldering fire test, but it is not clear that these fires are well represented by a lit cigarette applied to places where discarded cigarettes tend to land. The 31% of deaths associated with some kind of flaming ignition are numerous enough to justify our attention.

The question finally is what to do with this information. NFPA has been asked to develop a flaming-ignition test for upholstered furniture.  In the process, we are considering not just small flaming ignitions but also ignitions by another burning object.  Our goal is to be able to assess candidate technologies and designs for their effectiveness in addressing scenarios of importance.  The question of when and whether to build binding requirements around any test is a separate question that involves other considerations.

I don’t know anyone who wants to adopt a requirement that can be satisfied only by technologies that create more problems than they solve.  I understand different parties have different opinions about the implications of the evidence for reliability, effectiveness, cost, and non-fire effects.  I have a problem with anyone who wants to prejudge the debate by excluding certain fires or certain non-fire effects from the discussion.  I know my employer, NFPA, is trying very hard to be a fair and honest broker in these discussions and to keep the discussion and the search for solutions comprehensive in every sense of the word.  I hope readers of this piece will do likewise.

You can read additional articles on this topic published in the IAFSS Newsletter, Fire Safety Science News on this topic:

In Fire Safety Science News #36:
Effectiveness vs. Toxicity of Flame Retardants by David Rich
Open Flame Testing of Upholstered Furniture and Fire Safety by Marcelo M. Hirschler