Safety Experts Question Effectiveness of Popular Smoke Alarms
National News • July 24, 2006
When a jury this spring concluded a smoke alarm failed in a fatal upstate New York house fire, safety experts were already questioning whether popular models meet the threat posed by fast-burning ...
Insurance Journal is not responsible for the content of the message below.
Subject: Misleading Info from US Fire Administration Risking Lives
Posted On: August 15, 2006, 11:53 am CDT
Posted By: Joseph Fleming
Comment:
USFA Releases Report on Residential Structure Fire Fatalities
Aug 10, 2006, EMMITSBURG, Md. -- The U.S. Fire Administration (USFA) has released a special report regarding fatal residential structure fires with working smoke alarms. From 2001-2004, 391 fatal residential structure fires with working smoke alarms were reported to the National Fire Incident Reporting System (NFIRS), resulting in 452 civilian fire fatalities, according to the report.
"Each year an unacceptable number of American citizens lose their lives in residential structure fires where smoke alarms operated," says Acting USFA Fire Administrator Charlie Dickinson. "USFA has investigated the characteristics of the victims of these fires to better educate the public in taking preventative measures to reduce the risks involved with such fires."
According to the report, alarms operated in 34 percent of fatal apartment fires and in 12 percent of fatal one- and two-family dwelling fires. Thirty-seven percent of the victims of fatal residential structure fires with working smoke alarms were sleeping at the time of their death. An additional 29 percent of victims were trying to escape the fire at the time of their fatal injury.
The report, Investigation of Fatal Residential Structure Fires With Operational Smoke Alarms, was developed by the USFA's National Fire Data Center as part of its Topical Fire Research Series. The investigation results are based on data from the NFIRS for 2001-2004. The report examined residential structure fires with fatalities where smoke alarms operated, and addresses the characteristics of the fatalities.
"We know the value that smoke alarms, escape plans, child fire prevention programs and residential sprinklers can have on saving lives," says Dickinson. "We continue to encourage all residents of this nation to take the steps necessary to ensure their homes are fire safe today."
J. Fleming's Comments on Report
I would like to make the following comments regarding the USFA Report on smoke detectors.
From the Press Release
The U.S. Fire Administration has released a special report regarding fatal residential structure fires with working smoke alarms. The report reveals alarms operated in 34 percent of fatal apartment fires and in 12 percent of fatal one- and two-family dwelling fires.
But the report also states the following:
"Alarms were present in approximately 60 percent of fatal residential structure fires. Of those fatal fires with an alarm present, the detector operated 39 percent of the time. Together, these statistics indicate that alarms were present and operated in 23 percent of fatal residential structure fires".
Comment on the Validity of Either Statement.
In my opinion, the error in the press release quote posting stems from the confusing manner in which the USFA reports the data. The report does state in Figure 1 that in 1-2 family fatal fires only 12% had operational alarms. However the source of this data is never explained. In addition, the contradiction between this statement with other statement, which claims that "that alarms were present and operated in 23 percent of fatal residential structure fires", is never resolved.
But perhaps more confusing than the contradictory statement put forth by the USFA is that in my opinion neither the 12% figure nor the 23% figure is correct. I suspect it is closer to 40%.
Compare these 2 statements.
From "Fatal Fires" Topical Research Series Vol. 5 – Issue 1, USFA March 2005.
"Smoke alarms were present in 58% of fatal residential structure fires, but only operational in 37% of those fatal fires."
In this report the USFA meant the following:
- No Detector Present – 43%
- Detector Present and operated – 37%
- Detector Present and did not Operate – 21%
From "Fatal Fires with Operational Smoke Alarms" Topical Research Series Vol. 6 – Issue 2, USFA August 2006.
"Alarms were present in 60% of fatal residential structure fires. Of those fatal fires with an alarm present, the detector operated 39% of the time. Together these statistics indicate that alarms were present and operated in 23% of fatal residential structure fires."
The first 2 sentences, of this claim are very similar to the previous claim but in this report the third sentence provides a new and very different analysis. In this report the USFA appears to mean the following:
- No Detector Present – 40%
- Detector Present and operated – 23%
- Detector Present and did not Operate – 37%
Therefore, according to the USFA, from March of 2005 to August of 2006, the % of fatalities that occurred with an operating detector went from 37% down to 23%. While simultaneously, the % of fatalities that occurred with a detector that did not operate went from 21% to 37%. This makes no sense. I suspect that the author of this report mistakenly multiplied the 60% by 39% instead of leaving the 39% alone.
I suppose this could be due to a grammatical error except for the fact that it varied so much from previous estimates. Here is a summary of statistics from previous USFA Reports titled "Fire in the U.S." Increase in fatal fires with working detectors – 1988 (9%), 1994 (19%), 1996 (21%), 1998 (29%), 2001 (39%). While there will always be a certain percentage of people who cannot be saved by smoke detectors, e.g. the handicapped, those intimate with the fire etc., there is no reason to believe that the number of those people quadrupled between 1988 and 2001. In addition, while the number of fires with working detectors increase approximately in proportion to the increase in the number of detectors installed, the increase in the % of fatal fires with working detectors far exceeds it. Although this information was provided to the USFA, the USFA never made this disturbing increase a part of this recent study. Despite repeatedly stating in reports "In some cases the detector may have gone off too late to help the victim, or the victim may have been too incapacitated to react. But the % of deaths with detectors, especially the upward trend, is somewhat disturbing since there is a widespread belief that an operating detector will save lives. Further study is needed to show what other factors were involved with these deaths."
After repeatedly stating that this phenomenon should be studied the USFA failed to look at this issue in this paper. I will admit they spent a lot of time looking at why some people die when the detector works but they did no analysis whatsoever for why that number would have increased so much.
In my opinion the best explanation is that starting in the late 80's, Underwriters Labs, in an attempt to reduce nuisance alarms, forced the manufacturers to make less sensitive ionization detectors. Shortly after this change UL modified the smoldering test in UL217, the UL Smoke Detector Standard, in ways that made it much easier for the ionization detector to pass. The gradual introduction of these desensitized ionization detectors into American homes is the most reasonable explanation for the increase in fire deaths in cases where the detector operated. As a consequence this unrecognized flaw could be responsible for hundreds of deaths per year.
Another problem not really investigated by the Report is the number of fatalities that occur when the detector is disabled. Even if I am correct that the number of fatalities occurring when the detector is disabled is approximately 20% and not 37%, it is still a significant problem. If that is the case why does the USFA state, "Some models sound false alarms when they detect cooking smoke…". Why not point out that the models they are referring to are ionization detectors. Here is a quote from a recent study in Alaska.
"In this study they found that at the end of 6 months 19% of the homes with ionization detectors had disabled the detector and over 80% of the time the reason was that "it goes off too much" with 93% of the false alarms related to cooking. Only 4% of the photoelectric detectors were disables and none of the reasons were related to nuisance alarms."(Ionization and photoelectric smoke alarms in rural Alaskan homes", Alaska Injury Prevention Center, August 2000)
The USFA report recommends, "relocating alarms farther from sources of fumes," but what is someone to do in a small apartment or mobile home. Since HUD allows ionization detectors to be installed in mobile homes, where the layout guarantees they will be near a kitchen, we would expect to see a problem and we do. According to the USFA over 80% of the fatalities in mobile homes, where the operation of the detector is known, occurred when the detector was disabled. Why blame the victims in these cases? Why isn't HUD mandating photoelectric detectors for mobile homes? Why isn't the USFA sharing this information with the public?
I agree with both previous comments that it is desirable to hard-wire smoke detectors and as Fire Marshal for the Boston Fire Department I pushed for language encouraging this type of upgrade during renovations. However I think that hard-wiring ionization detectors is not part of the solution but part of the problem.
Subject: Misleading Info from US Fire Administration Risking Lives
Aug 10, 2006, EMMITSBURG, Md. -- The U.S. Fire Administration (USFA) has released a special report regarding fatal residential structure fires with working smoke alarms. From 2001-2004, 391 fatal residential structure fires with working smoke alarms were reported to the National Fire Incident Reporting System (NFIRS), resulting in 452 civilian fire fatalities, according to the report.
"Each year an unacceptable number of American citizens lose their lives in residential structure fires where smoke alarms operated," says Acting USFA Fire Administrator Charlie Dickinson. "USFA has investigated the characteristics of the victims of these fires to better educate the public in taking preventative measures to reduce the risks involved with such fires."
According to the report, alarms operated in 34 percent of fatal apartment fires and in 12 percent of fatal one- and two-family dwelling fires. Thirty-seven percent of the victims of fatal residential structure fires with working smoke alarms were sleeping at the time of their death. An additional 29 percent of victims were trying to escape the fire at the time of their fatal injury.
The report, Investigation of Fatal Residential Structure Fires With Operational Smoke Alarms, was developed by the USFA's National Fire Data Center as part of its Topical Fire Research Series. The investigation results are based on data from the NFIRS for 2001-2004. The report examined residential structure fires with fatalities where smoke alarms operated, and addresses the characteristics of the fatalities.
"We know the value that smoke alarms, escape plans, child fire prevention programs and residential sprinklers can have on saving lives," says Dickinson. "We continue to encourage all residents of this nation to take the steps necessary to ensure their homes are fire safe today."
J. Fleming's Comments on Report
I would like to make the following comments regarding the USFA Report on smoke detectors.
From the Press Release
The U.S. Fire Administration has released a special report regarding fatal residential structure fires with working smoke alarms. The report reveals alarms operated in 34 percent of fatal apartment fires and in 12 percent of fatal one- and two-family dwelling fires.
But the report also states the following:
"Alarms were present in approximately 60 percent of fatal residential structure fires. Of those fatal fires with an alarm present, the detector operated 39 percent of the time. Together, these statistics indicate that alarms were present and operated in 23 percent of fatal residential structure fires".
Comment on the Validity of Either Statement.
In my opinion, the error in the press release quote posting stems from the confusing manner in which the USFA reports the data. The report does state in Figure 1 that in 1-2 family fatal fires only 12% had operational alarms. However the source of this data is never explained. In addition, the contradiction between this statement with other statement, which claims that "that alarms were present and operated in 23 percent of fatal residential structure fires", is never resolved.
But perhaps more confusing than the contradictory statement put forth by the USFA is that in my opinion neither the 12% figure nor the 23% figure is correct. I suspect it is closer to 40%.
Compare these 2 statements.
From "Fatal Fires" Topical Research Series Vol. 5 – Issue 1, USFA March 2005.
"Smoke alarms were present in 58% of fatal residential structure fires, but only operational in 37% of those fatal fires."
In this report the USFA meant the following:
- No Detector Present – 43%
- Detector Present and operated – 37%
- Detector Present and did not Operate – 21%
From "Fatal Fires with Operational Smoke Alarms" Topical Research Series Vol. 6 – Issue 2, USFA August 2006.
"Alarms were present in 60% of fatal residential structure fires. Of those fatal fires with an alarm present, the detector operated 39% of the time. Together these statistics indicate that alarms were present and operated in 23% of fatal residential structure fires."
The first 2 sentences, of this claim are very similar to the previous claim but in this report the third sentence provides a new and very different analysis. In this report the USFA appears to mean the following:
- No Detector Present – 40%
- Detector Present and operated – 23%
- Detector Present and did not Operate – 37%
Therefore, according to the USFA, from March of 2005 to August of 2006, the % of fatalities that occurred with an operating detector went from 37% down to 23%. While simultaneously, the % of fatalities that occurred with a detector that did not operate went from 21% to 37%. This makes no sense. I suspect that the author of this report mistakenly multiplied the 60% by 39% instead of leaving the 39% alone.
I suppose this could be due to a grammatical error except for the fact that it varied so much from previous estimates. Here is a summary of statistics from previous USFA Reports titled "Fire in the U.S." Increase in fatal fires with working detectors – 1988 (9%), 1994 (19%), 1996 (21%), 1998 (29%), 2001 (39%). While there will always be a certain percentage of people who cannot be saved by smoke detectors, e.g. the handicapped, those intimate with the fire etc., there is no reason to believe that the number of those people quadrupled between 1988 and 2001. In addition, while the number of fires with working detectors increase approximately in proportion to the increase in the number of detectors installed, the increase in the % of fatal fires with working detectors far exceeds it. Although this information was provided to the USFA, the USFA never made this disturbing increase a part of this recent study. Despite repeatedly stating in reports "In some cases the detector may have gone off too late to help the victim, or the victim may have been too incapacitated to react. But the % of deaths with detectors, especially the upward trend, is somewhat disturbing since there is a widespread belief that an operating detector will save lives. Further study is needed to show what other factors were involved with these deaths."
After repeatedly stating that this phenomenon should be studied the USFA failed to look at this issue in this paper. I will admit they spent a lot of time looking at why some people die when the detector works but they did no analysis whatsoever for why that number would have increased so much.
In my opinion the best explanation is that starting in the late 80's, Underwriters Labs, in an attempt to reduce nuisance alarms, forced the manufacturers to make less sensitive ionization detectors. Shortly after this change UL modified the smoldering test in UL217, the UL Smoke Detector Standard, in ways that made it much easier for the ionization detector to pass. The gradual introduction of these desensitized ionization detectors into American homes is the most reasonable explanation for the increase in fire deaths in cases where the detector operated. As a consequence this unrecognized flaw could be responsible for hundreds of deaths per year.
Another problem not really investigated by the Report is the number of fatalities that occur when the detector is disabled. Even if I am correct that the number of fatalities occurring when the detector is disabled is approximately 20% and not 37%, it is still a significant problem. If that is the case why does the USFA state, "Some models sound false alarms when they detect cooking smoke…". Why not point out that the models they are referring to are ionization detectors. Here is a quote from a recent study in Alaska.
"In this study they found that at the end of 6 months 19% of the homes with ionization detectors had disabled the detector and over 80% of the time the reason was that "it goes off too much" with 93% of the false alarms related to cooking. Only 4% of the photoelectric detectors were disables and none of the reasons were related to nuisance alarms."(Ionization and photoelectric smoke alarms in rural Alaskan homes", Alaska Injury Prevention Center, August 2000)
The USFA report recommends, "relocating alarms farther from sources of fumes," but what is someone to do in a small apartment or mobile home. Since HUD allows ionization detectors to be installed in mobile homes, where the layout guarantees they will be near a kitchen, we would expect to see a problem and we do. According to the USFA over 80% of the fatalities in mobile homes, where the operation of the detector is known, occurred when the detector was disabled. Why blame the victims in these cases? Why isn't HUD mandating photoelectric detectors for mobile homes? Why isn't the USFA sharing this information with the public?
I agree with both previous comments that it is desirable to hard-wire smoke detectors and as Fire Marshal for the Boston Fire Department I pushed for language encouraging this type of upgrade during renovations. However I think that hard-wiring ionization detectors is not part of the solution but part of the problem.