Why Smoke Detectors Don’t Fulfill Great Expectations for Saving Lives
Fire Prevention in Action
“A Detector in Every Home Will Save Lives.”
“Detection program Dramatically Reduces Fire Damage.”
“So go the headlines. And yet, speaking at the 1979 Fire and Life Safety Conference, Gordon Vickery of the United States Fire Administration had this sobering thought: “We put 50 million smoke detectors in buildings in America in a two-year period, and our fire loss and death loss rate goes up. We’re having a little trouble explaining these things …”
At a National Fire Academy gathering of state fire training directors in February 1980, another USFA official said fire deaths are “still on the upswing.”
According to a 1973 UL estimate, about 18,000 households would need detector protection to bring about the saving of one life per year. To really cut the total fire death toll, then, would require detectors in every home.
It is true that at least 28 states now mandate detector installation (though many don’t include single-family dwellings). Strictly local laws exist in 18 states. New homes—including mobile homes—with FHA and HUD backing must have detectors. Tens of millions of detectors have been sold in recent years.
So we should be seeing a drop in annual fire deaths. But that hasn’t happened. Do smoke detectors save lives? Does a fire department reduce fire losses? The answer to both questions is the same: “It depends on how they’re used.”
Unfortunately, just as a citizen may assume his home is safe because there’s a fire station on the next corner, so a resident may assume he’s fully protected because he’s purchased a smoke detector. Such devices are only a small part of the solution to the problem of public carelessness with fire. Undue stress on detector installation as the ultimate in household fire protection may well be lulling the public into a sense of false security.
Not that the devices don’t work. Far from it. Smoke detector technology is mature. That is, although new developments can be expected, the basic design and manufacturing breakthroughs have been made. Reliable manufacturers are in the field. National standards exist.
In becoming mature, residential alarm technology has survived a series of controversies. First was the argument between heat sensors and smoke sensors—still not fully resolved, although the smoke detector seems to have emerged victorious.
Second was the argument between ionization detectors (most responsive to free-burning fires) and the photoelectric type (more sensitive to smoldering, smoky blazes that represent 75 percent of building fires, according to studies by the Los Angeles Fire Department in 1963). Each has its place. Some detectors now combine both types in a single unit.
The most recent detector controversy has concerned radioactivity emitted by the ionization type, and the public has been amply reassured on that score.
But whether or not detectors work effectively, and keep on working, it is up to the user—not the manufacturer.
Apathy of homeowner
In seeking legislation to compel home detector installation, authorities complained 15 years ago that “generally the homeowner seems apathetic.” Such legislation is more widespread now; is the apathy any less? Will an apathetic resident keep in working order, and be prepared to respond to, the detector he installed only because it was the law?
In a 1979 report, New York University Medical Center researchers offered the reminder that “people do not always realize that their lives are in danger when warned of a fire, and do not always behave accordingly…”
Besides, it is in the low-income inner city areas where most fire fatalities occur. Residents there have trouble enough reaching the landlord or the property manager for corrections of plumbing, heating, or electrical problems. Will they be more likely to get prompt attention if their smoke detectors need maintenance? Investigation of some fatal fires has disclosed that detector batteries had been removed because they were so often stolen. Few residents own their homes. Few are fire safety conscious—and not likely to be made so by installation of a detector they know nothing about.
Detector limitations cited
Smoke detectors in every home cannot strongly influence residential fire safety unless users and the fire service understand their limitations. This was brought out in a recent report by the University of Toledo, concluding that:
- Legislation requiring the installation of smoke detectors appears to be accompanied by complacency.
- Very few families were aware of their need, with or without detectors, for a well-planned and practiced exit in case of fire.
A hard sell, rather than new laws, has been the fire service tactic in some places. “If it can sell toothpaste, it surely will sell smoke detectors” was the approach a West Coast fire department took in setting up radio commercials. This promotional campaign sent detector sales soaring throughout the area.
But is that really building a high level of home fire safety?
A detector manufacturer recently offered the public commercials showing a family casually exiting the second floor of their home down a smoky stairway after being alerted to a fire on the first floor. The Seattle Fire Department, for one, has expressed concern over the image conveyed by such advertising, or by another commercial showing flaming newspapers being used to test a household detector.
Another West Coast city recently reported on a program of installing free smoke detectors in the homes of elderly residents. Recognizing the importance of maintenance, the fire department first installed the units, then set up a continuing program of inspection visits to check their operation. But how many fire departments can undertake such a program, which needed a state grant to get under way?
During 1977 in an eastern city, community development funds were used to give smoke detectors to 2000 residents. It was the city’s goal to eventually provide every family with at least one detector. Such programs, however, may yield only short-term benefits. Keeping those households safe will require much more than simply having detection on the premises. (A third conclusion of the University of Toledo study was that “people who buy smoke detectors voluntarily often install them incorrectly.”)
In five years or 10, will detectors remain in use and in good order? Though there are many stories of lives saved when residents were aroused by household alarms, there are also stories of detectors without batteries, or with disconnected wiring, or otherwise inoperable for lack of maintenance.
In its helpful booklet, “What You Should Know About Smoke Detectors,” the U. S. Consumer Product Safety Commission deals with the common householder question, “How do we take care of it?” Monthly testing with real smoke is advised for older models, but the booklet adds, “Check the package for instructions of your detector to see if it has this feature.” Neither a renter, nor the second or third owner of the property, is likely to have either package or instructions.
NFPA 74 requires that the detector buyer be given testing and maintenance instructions, a guide for setting up a home evacuation plan, and sources of repair parts and replacement service. But it’s up to the homeowner to keep that information handy, to keep it current as repair firms move or quit business, to act on the information, and to pass it on to subsequent residents. In many suburban neighborhoods, ownership turnover is 10 or 20 percent annually. A home may change tenants every three or four years.
Batteries should be replaced as soon as the low-power warning sounds. Unfortunately, several widely used detector models require a special battery that is often hard to find. Keeping an extra one on hand is a good idea, but there will be no “low-power warning” for the spare sitting on a closet shelf. The National Fire Prevention and Control Administration pointed out in 1977 that codes requiring home detectors often specified AC-powered, plug-in units because “field data indicate that many purchasers …. have not replaced wornout batteries immediately . . People who did not voluntarily purchase detectors are expected to be casual about battery replacement.”
Two years ago, an electronics technician in Texas, despite his specialized knowledge and long experience, accidentally put in a new 9-volt detector battery incorrectly and blew out part of the unit’s circuitry. Not all detector makes are arranged to prevent such accidents.
Moreover, the warning signal is normally activated by decreasing battery terminal voltage as cells get older. But that is only one way in which batteries fail. Internal resistance rises as a battery deteriorates, eventually making it impossible for the battery to supply enough current to operate the alarm even though no-load battery terminal voltage remains high. Researchers concluded in 1978 that about 3 percent of detector batteries were then becoming unusable before the low-power warning was sounded. Only frequent checks of detector operation will reveal this.
Detectors using house current instead of batteries avoid such problems. But user complacency can be fatal with this type as well. Here’s a case history:
“Because they had both heat detectors and a smoke detector installed in their home, this… family probably felt safe from fire. However, the heat detection system had been disconnected temporarily and, when fire struck, power to the smoke detector failed … Three residents died in what was a relatively minor fire.”
Further maintenance needed for some types of detectors includes periodic cleaning of the smoke chamber, where accumulated dust can trigger false alarms. Users may not realize the number of such alarms that can occur, ranging from 10 to 50 times the number of actual fire responses. The first major study of this problem was in a new Texas community, The Woodlands, after three years of experience with what grew to 1100 protected homes by 1978. During those three years, there were nearly 1500 false fire alarms from detectors in that community. Most of them originated in kitchen areas.
The capability of detectors to respond to actual fires has been established by several major field tests. These have been aimed at finding out what specific fire, smoke and heat buildup exists in various parts of a structure at the time of detector operation. However, a crucial question that has been widely ignored is, “Will that detector waken a sleeping occupant in time?”
The first big test was in 1960 by the engineering unit of the Los Angeles Fire Department. From 13 blazes set in a variety of dwellings, it was found that heat detectors worked fastest in freely burning fires, whereas smoke detectors had the edge in smoldering conditions.
Lengthy arguments about such testing have involved such questions as: What concentration of CO or smoke is deadly? How does that vary with a person’s age or health, or with the particular kind of smoke? Has it changed with the use of new plastics in the home? How typical of real dwellings were the contents used in test fires?
Such arguments intensified after the famous 1974-76 Indiana Dunes tests. These repeated in far greater detail the earlier Los Angeles work. The smoke detector now appeared more generally useful, so the Dunes tests concentrated on this type—which added to the controversy.
An important new concept appeared in this testing—the idea of escape time. Based on both human and animal reactions to fire gas exposure in the laboratory, conditions in the test dwellings when detectors operated were compared to those under which an occupant would presumably become unable to navigate his way out. That permitted an estimate of the individual’s maximum available time to escape from various parts of the building after detectors operated.
But little attention seems to have been paid to the more basic question: Is it valid to assume the occupant’s try to escape will commence the instant the alarm sounds?
This question was dealt with in the most recent (1978) set of major tests, which again took place in Los Angeles. To start with, detectors were installed in two vacant dwellings containing typical furnishings. Eventually, 1200 off-theshelf smoke and heat alarms were used in 71 separate fire tests to produce over 2 million computer-analyzed pieces of data, resulting in a draft report 2’/2 inches thick.
As this is written, those California tests are reported to be “quite controversial,” their validity questioned by some scientists. Copies are not available until revisions of the draft can be completed. A source close to the project states that the final report “intends to allow the reader to arrive at his own conclusion . . . The USFA may be sponsoring publication.”
This work began as a direct outgrowth of some of the disagreement about the Dunes program. Joint sponsors were the California Fire Chiefs Association, the LA City Fire Department, and a number of detector manufacturers. Many other agencies, firms, schools and individuals took part. To remove bias from the results, all the generic detector types on the California market were included, each identified only by a code number with all other labels or markings removed.. The types of fires themselves were chosen from fire incident report statistics of the State Fire Marshal’s Office as being the most common kinds of residential fires to be expected.
To shed more light on escape time, some non-fire tests were added. According to the report draft, here’s how these worked:
“A number of test boxes were constructed which contained a digital clock, a timer and a detector. The components were connected so that when the box was plugged into an electrical outlet, the timer was activated. After a pre-set interval had elapsed (usually two to four hours), the alarm was activated simultaneously with the alarm and clock were stopped by disconnecting the electrical cord from the wall outlet.
“Both fire personnel and civilians volunteered to participate. Each household was instructed to place the box in a central location near the bedrooms and the person going to bed last was to plug in the box, thus beginning the timer. During the night, the alarm would activate at the pre-set time and all persons in the house were to evacuate and assemble at a prearranged location.
The last person to leave the house was supposed to make sure that all persons were out of the house and then to disconnect the box … When the box was returned the next day it was opened and the elapsed time was recorded, providing an estimate of escape time.”
Weaknesses in testing
This is important data. However, the procedure contains several unavoidable weaknesses. First, those “escaping” were unhindered by smoke, gas, heat, or the terror of suddenly finding themselves and their families in mortal danger. Second, though they didn’t know when the alarm would sound, they did know that it eventually would. Anticipation tends to speed response. Third, each volunteer was given the EDITH exit plan to follow—excellent training in any household, but not yet typical of the general population.
Another useful new addition to the testing was data on detector audibility. Sound level is measured in units called decibels, or dB. The recent California tests showed that (from a number of hallway detectors) the average alarm sound raised the preexisting sound level in a bedroom with closed door by 19.6 dB. If the door were open, room noise level went up by about 30 dB when the alarm sounded.
In an average bedroom at night, the existing, or ambient, sound level is about 25 dB—see figure 1. That means the alarm in the hall outside the bedroom produces room noise at 45 to 55 dB, depending on whether the door is open or shut.
Will that wake sleepers within the room? The appendix to NFPA 74, the standard for residential detectors, assumed bedroom ambient at 55 dB and recommends minimum sound output from detector alarms of 70 dB within the room to be sure sleepers will be aroused. From the California tests, it seems doubtful that standard detectors will raise the room noise level that high.
Much more doubt, however, arises from other scientific data on the soundness of sleep and how much noise it ‘.akes to wake people. This comes largely from one researcher—Dr. Michael Bonnet of the Sleep Laboratory of the University of Cincinnati. Bonnet has spent several years studying the five basic stages of sleep and the difficulty of arousing persons from each stage.
Says Bonnet, “While considerable evidence has accumulated demonstrating the effectiveness of the smoke detectors in detecting smoke no evidence exists that they can indeed alert ‘even the soundest sleeper’ as claimed by detector advertisements …. Experimental evidence indicates that thresholds in stage 4 sleep range between 60 and 120 dB, and smoke alarms produce an 85-dB signal which is decreased a further 15 dB by passage through a closed bedroom door.” (The California tests indicate the decrease may be greater.)
In an experiment with 10 young adults, a smoke alarm was set off for 30 to 60 second intervals at some random time during the night. Bonnet found that the subjects were awakened only 42 percent of the time. Besides the loudness of the alarm itself, the sleeper’s expectations and the nature of the alarm sound also contribute to its effectiveness, he contends, adding, “An alarm clock goes off when you expect it to and it’s also placed near your bed. However, a smoke alarm is usually located at some distance from the sleeper and is easy to be ‘ignored’—especially its level, monotonous tone.”
That comment about the type of alarm sound recalls the 1973-75 argument about the most effective kind of standard alarm tone in public buildings. There was strong support for what was called a slow whoop tone, a rising and falling non-uniform sound. To get more complex sounds from a home fire detector is an added design and cost burden, yet the literature indicates that little testing has been done to establish what the best home alarm sound is.
Consumers Union detector tests in 1976 reported actual sound levels of 82 to 90 dB, varying in loudness from sample to sample of the same make. Some detectors first responded by “tiny hiccups, or a brief activation of the alarm,” going into the full steady sound only after several minutes. One make did have a “pulsating alarm judged likelier than most to awaken a deep sleeper.”
Effectiveness of sounds
Recently, a safety director at the Massachusetts Institute of Technology asked, “What is the relative effectiveness of horns, buzzers, pitch tones, and bells as devices to wake sleeping persons? Is loudness the only consideration?”
Several years ago, dormitory residents complained to his office that building fire alarms were not loud enough to wake them. Tests then showed that alarm clocks or clock radios were effective, with measured sound level at the pillow of 55 to 60 dB. So the building alarms were supplemented to provide that sound level.
More recently, however, complaints of insufficient alarm loudness are again being received. Are young people’s ears becoming less sensitive after increasing assault by rock and disco sound? No one knows.
Moreover, Bonnet has pointed out that sleep is much deeper in the first half of the night. The use of sleeping pills or some other medications can increase the depth of sleep.
“I remember instances in my experiments,” he added, “where a 115-dB tone did not awaken a subject, and where entering the room and speaking his name did not awaken him. In some cases, it has been necessary to shake normal adults to wake them up. No detector owner or user should be unaware of this. I believe that further research is very important . . . and that standards for probability of arousal be set and enforced based upon that research.”
Sleeps through shootings
In Milwaukee last February, three occupants of a small home were shot to death, while a fourth person asleep in the house heard nothing. In the April 1977 issue of Consumer Reports, a homeowner described how a smoky fireplace set off a detector between his children’s bedrooms, while they slept with bedroom doors open, and he was awake. Much to his surprise, the alarm awakened neither of the children.
Detector industry spokesmen themselves said the incident “raised serious questions about the adequacy of NFPA Standard 74’s minimum audibility requirement of 85 dB.” They asked the National Fire Prevention and Control Administration to “investigate known data on the minimum sound level necessary to awaken an ‘average’ person,” and also to “perform a test series to determine what sound levels must be generated at remote locations to meet this minimum … in the bedroom.”
Writing two years ago in the NFPA Fire Journal under the title, “Will Your Smoke Detector Wake You?” Charles Berry cited other studies showing that it took 75 dB to waken young adults as much as half the time. In still other tests, it took between 72 and 90 dB— and at least 25 dB more if sleeping pills had been used.
Turn to page 57
His conclusion: “Persons with hearing impairments, those taking sleeping medicine, and those who regularly ‘have a couple of nightcaps’ should consider that 75 dB will probably not be loud enough to awaken them and, in fact, they may require over 100 dB…. The homeowner must determine experimentally whether he personally can be awakened by a detector located some distance from himself. Few homeowners are likely to go to the trouble.”
Berry proposed further research, recommending meantime that detector standards adopt what he called an “awakening worksheet” as part of NFPA 74, a copy to be packaged with each detector (see Figure 2). This sheet permits calculation of the probability of awakening when a detector alarm sounds. Based on many assumptions, including average hearing capability and sleep thresholds, this sheet is not a guarantee—only a guide.
In the Dunes tests, many of the escape time calculations, as in figure 3, showed that most people would have only a minute or two to escape a fire once detectors had operated. Some persons might be unable to escape at all. And this presumes that escape attempts begin as soon as the alarm sounds. The more recent sleep research indicates that the available escape time could all be gone before the sleeper is aroused.
From the evidence, contends Bonnet, “it must be concluded that safety claims concerning arousal made for presentday smoke alarms are probably unfounded and that the implied sense of security given by them is questionable.”
Advice for detector users
The Better Sleep Council makes these recommendations for home fire detector users:
- Realize the limitations of a smoke alarm and do everything to eliminate dangerous conditions around the house that might result in a fire.
- Locate the alarm nearer to adults than children, since children tend to be deeper sleepers. If there is a very light sleeper in the house, put the alarm closest to that person. At the same time, remember the need to have a detector near the likeliest area of fire occurrence.
- An expensive but effective way to increase the efficiency of an alarm system is to have several alarms located in different areas, wired in such a way that when one goes off, all of them do. (Such interconnection is also suggested in NFPA 74.)
- Try to have a few test sessions where you can arrange through friends or relatives to have your alarm activated late at night when you don’t expect it.
According to Dr. A. W. Phillips of the National Smoke, Fire and Burn Foundation, Inc., studies of human behavior in fires recently showed that irrational or uncoordinated behavior resulting from Fire gases is aggravated by preexisting illness or drinking. In Maryland, half the fatalities from “smoke” involved victims with significant blood alcohol levels. Phillips claims that, once aware of fire danger, “what the victim does, if he or she is able to act, may depend on his or her physical and emotional Fitness, education and training, past fire experience, the presence or absence of loved ones … and his or her intelligence and sense of responsibility.” Clearly, the alarm itself is only one of many ingredients in a successful escape.
In 1978, the NFPA studied 64 cases of smoke detector performance in dwelling fires since 1974. The detectors did not work successfully in 23 of these cases. Improper location caused three failures; removal or poor maintenance produced seven malfunctions; “arousal or evacuation problems” caused 12. Some victims had been drinking. At least one was deaf. The conclusion was that “evacuation planning must be emphasized in promoting the installation of smoke detectors in residential occupancies. In addition, special problems such as hearing impairments must be considered in fire safety planning.”
Here are two examples of such failures:
In Massachusetts, four residents of a two-story dwelling died although a smoke detector on the ceiling of one bedroom did operate and was still sounding when fire fighters arrived. Investigators felt the noise of an air conditioner also operating in the same room may have been partly responsible.
In Virginia, a smoldering furniture fire set off a smoke detector in an apartment about 5 a.m. Its occupant did not wake up. But someone in the apartment upstairs did hear the alarm. Another then entered the apartment of origin to find the occupant still asleep, but was able to rouse him in time.
Exterior alarm considered
Earlier this year the NFPA agreed to consider a proposal that Standard 74 require “an outside bell or other audible device” to be operated by detectors in a household with handicapped or invalid occupants. Infants, intoxicated or “medicated” persons, those who smoke in bed, and those “unpracticed or unknowledgeable in proper escape procedures” would thus have a chance for rescue by neighbors or passersby outside the home—presuming such rescuers would be at hand when needed.
Various United States and Canadian studies have concluded that smoke alarms could save between 25 and 66 percent of all fire victims. All such figures rest on three foundations: (1) that the detectors are located correctly and remain in working order; (2) that occupants are immediately aroused by an alarm; and (3) that once aroused, each person has the physical and mental capacity to escape. There is considerable evidence that one or more of those presumptions is incorrect much of the time.
As the NFPA has pointed out, despite its strong support for home detectors, “detection equipment neither prevents Fires nor puts Fires out. All that detection equipment can do is warn occupants that there is a fire and they had better leave the building—fast. If the occupants are unresponsive, if they don’t know how to react in a Fire, if they have not developed an escape plan, the detection equipment may do them little good.”