Contents
The National Fire Incident Reporting System (NFIRS) is the standard system in which fire departments in the United States report their incidents, so that reports can be combined into state and national databases.[1] Ahrens and Stewart (2008) describe it as the largest and most detailed fire incident database in the world and the source of most of what is known about fire in the United States.[1] The National Fire Data Center set it up after the National Fire Prevention and Control Act of 1974, and its classifications grew out of the coding standard NFPA 901.[2]
NFIRS has been widely used since 1980. Version 5.0, first introduced in a few areas in 1999, is a modular system designed to capture all types of incidents, not just fires.[1][3][4] The database is not a census of fires: the 2008 NFPA handbook chapters put its share at one-third to two-thirds of the fires reported to local fire departments.[5][6] National totals are therefore estimated by scaling NFIRS data with the results of a separate sample survey run by the National Fire Protection Association.[7] This article describes the system as the 2008 NFPA handbook and the 2002 SFPE handbook present it.
History and purpose
In the early 1970s, even the best fire departments typically kept their fire incident records only as narratives, filed by incident number. Very few states had any fire incident database.[8] Extracting statistics from narrative reports was frustrating at local level and nearly impossible at state and national level.[3] National estimates came from the annual survey of the National Fire Protection Association (NFPA), and analyses of fire causes from its anecdotal Fire Incident Data Organization (FIDO) database.[8]
According to Ahrens and Stewart (2008), the need for aggregated fire data was identified in 1972, when President Nixon's Commission on Fire Prevention and Control issued its report America Burning. The report noted that the United States then had one of the highest fire death rates in the industrialised world.[1] It led directly to the National Fire Prevention and Control Act of 1974 (Public Law 93-498), which directed the formation of the U.S. Fire Administration, first called the National Fire Prevention and Control Administration.[1] Section 9 of the act set up a National Fire Data Center to "provide an accurate, nationwide analysis of the fire problem" and to "monitor the effectiveness of programs to reduce fire losses".[2][9] The National Fire Data Center established NFIRS to meet that charter.[3]
The coding came from NFPA. By 1969 NFPA had identified data elements and standardised classifications for recording fire incidents, and NFPA 901, Uniform Coding for Fire Protection, provided the classification system for the new NFIRS.[3] The 1976 edition of the incident and civilian casualty codes and the 1981 edition of the fire service casualty codes remained in use until NFIRS 5.0 was released in 1999.[3]
National coverage grew slowly. The first Fire in the United States report, published in December 1978, had a full year of detailed cause data for only two states, California and Ohio.[3] The second, published in 1982, covered 1978 fires with full-year data for 15 states. With federal funding, 38 states and the District of Columbia were participating or preparing to participate by April 1981.[3] Hall and Ahrens (2002) state that since 1980 some or all communities in roughly 40 of the 50 states have taken part.[10]
Structure and reporting process
NFIRS works on three tiers: local fire departments, state fire agencies and the federal government.[3] In 2004, more than 19,000 fire departments, over 60 percent of the national total, reported at least one incident to NFIRS.[4] Because Version 5.0 captures all incident types, a chief can see which companies are busiest and which fire problems lead. Some departments use the statistics to justify budgets, and applicants to the 2006 Assistance to Firefighters Grant Program had to participate in NFIRS.[4]
A completed fire report serves three purposes. It is a legal public record of the incident, and it informs senior officers and managers. It also supplies data on the fire problem for tracking trends and judging prevention and suppression measures.[3] The third purpose requires a standardised format, which NFIRS provides. Officers using it record the time of the incident, the occupancy, the area of origin, what caught fire, what started the fire, property damage and casualties. They also record whether sprinklers and detection equipment were present and whether they worked.[3]
At local level an NFIRS report is first an organised record-keeping system, and the reports are public records. Where the needs of researchers and of the fire service conflict, the needs of the fire service take priority.[4] The local fire department generally has the primary responsibility to document a fire and to make an initial determination of its origin and cause. A department that reports to NFIRS may also record factors that affected the course of the fire, such as flame spread and the performance of fire protection features.[11] Ahrens and Stewart (2008) treat cause and origin skills as an essential part of local NFIRS. Those who report should be trained and able to investigate origin, cause and circumstances, because weaknesses in fire investigation can undermine the validity of the data and of analyses based on them.[4]
States generally set the minimum data requirements for their departments. Many accepted only Version 5.0 data when the handbook was written.[12] After error checking, state program managers forward the data incident by incident to the federal server. Because the national system holds individual incidents rather than state totals, new questions can be put to the data later.[12] The National Fire Data Center processes the data and publishes Fire in the United States, and users such as NFPA and the U.S. Consumer Product Safety Commission (CPSC) obtain the data for their own analyses.[12]
Definitions and coding
A standardised system has to fix the meaning of everyday words. NFIRS counts only "hostile" fires. Its 5.0 reference guide defines a fire as "Any instance of destructive and uncontrolled burning, including explosion, of combustible solids, liquids, or gases".[12] Lightning, pressure-vessel ruptures, explosions of munitions, vehicle accidents and overheat conditions are excluded unless they cause a fire or result from one. Burning with a permit, and unauthorised burning, are not counted as fires unless they get out of control.[12]
Structures include buildings but also tents, storage tanks, tunnels and covered walkways.[12] Under Version 5.0 a vehicle burning inside a structure is a vehicle fire unless the structure itself becomes involved. Certain confined structure fires may be reported in abbreviated form with only the Basic Module. They include fires confined to cooking vessels, chimneys, incinerators, fuel burners or boilers, trash compactors, and trash with no flame damage to the structure or contents.[13]
Civilians are everyone who is not a member of the fire service, including police and other emergency personnel. An NFIRS fire death is anyone who dies as a result of the fire within one year of it.[13] Smoke inhalation and burns are the most common causes, but a person fatally injured while jumping from a burning building, or struck by a fatal heart attack during a fire, also counts. People who die immediately of trauma are not counted; for vehicle fires caused by collisions, autopsy results may be needed to decide.[13]
Version 5.0 introduced a broad cause field with the codes intentional, unintentional, failure of equipment or heat source, act of nature, cause under investigation, and cause undetermined after investigation.[14] "Intentional" replaced the older ignition factors "incendiary" and "suspicious" in the Fire Module. None of these terms corresponds exactly to arson. Dropping "suspicious", together with the lack of required cause coding for some incidents, was likely to produce an artificial drop in recorded arson.[13] With heat source, equipment involved in ignition, item first ignited, type of material ignited and factors contributing to ignition, the cause field gives NFIRS six cause-related data elements. Their combinations can define billions of fire cause categories, so analysts use grouped approaches.[15]
Version 5.0 and its modules
The redesign began in 1989, after feedback in the late 1980s showed that users wanted a sweeping revision of Version 4.1. The federal agency and the National Fire Information Council ran workgroups of departments, state managers, data users and software vendors over seven years.[14] Every data element had to be reportable, collectable and usable. Firefighters found it hard to judge reliably whether all openings were protected, so the field "construction type" was dropped.[14]
Version 5.0 is modular. The Basic Module records identifying and dispatch data for every incident, and the Fire Module is required for nonconfined fires.[14] The Structure Fire, Civilian Fire Casualty and Fire Service Casualty Modules are required when the incident or its casualties call for them.[14] The EMS, Hazmat, Wildland Fire, Arson, Juvenile Firesetter, Apparatus or Resources, Personnel and Supplemental Modules may or may not be required, or even accepted, at state or local level.[16]
The new version added detail that earlier versions lacked. Three-digit incident types separate false alarms and unintentional sprinkler activations, and the Structure Fire Module records the presence, type and effectiveness of sprinkler and detection systems and the reason for any failure.[14][17] The casualty modules grade severity from minor to death, and the Civilian Casualty Module records the victim's floor at ignition and at injury, details of interest for performance-based codes.[18] Direct property damage, one measure of the severity of fires and of fire loss, is estimated by the fire service on the basis of expected replacement cost, with separate fields for structure and contents.[17]
The transition took years. In 1999, 93 percent of the data were collected in Version 4.1 and 7 percent natively in Version 5.0; by 2003 the shares were 18 and 82 percent.[19] Older data are converted to the Version 5.0 format, but the conversion is not one-to-one because code definitions differ. The version used for collection is recorded, so analysts can include or exclude converted data.[19] The national database for 1999 through at least 2004 therefore mixes data from different versions.[20]
Coverage and national estimates
At national level NFIRS is neither a complete census of reported fires nor a random sample. It cannot by itself answer questions such as how many fires were reported, or how many people died in fires, in a given year.[21] Each state sets its own requirements: some require all runs to be reported, some only fires, some apply a dollar-loss threshold, and in some reporting is voluntary.[21]
In 2003, more than 18,000 fire departments reported more than 930,000 fires, 59 percent of the estimated total reported to local fire departments.[22] The same chapter also says that typically roughly half to two-thirds of all fires are represented, while Ahrens and Stewart (2008) give between one-third and one-half of the fires reported to municipal fire departments.[6][5] Hall (2002c) gives nearly a million incidents a year, typically close to half the total.[23] As an illustration, dividing the 930,000 NFIRS fires by NFPA's estimate of 1,584,500 fires in 2003 gives about 58.7 percent, consistent with the 59 percent quoted.[22][24]
With so many records, the sampling precision of NFIRS is outstanding. Because it is not a true random sample, however, its bias cannot be calculated by standard statistical methods.[23] Participation and reporting practices are not uniform across regions and community sizes, and some communities do not report fires with zero property damage. As region and community size are correlated with the frequency and severity of fires, NFIRS may be susceptible to systematic bias.[10] Hall and Ahrens (2002) judge that poorer communities are proportionally less well represented, although the differences appear modest.[25]
National estimates combine the detail of NFIRS with the totals of the NFPA fire experience survey, a stratified random sample of fire departments.[10] Analysts compare the survey totals with the NFIRS totals for the same category, and invert the captured fraction into a scaling ratio. Analysts need the two databases together to develop national estimates.[7]
| Symbol | Description | Unit |
|---|---|---|
| Scaling ratio for property class k and loss measure m | – | |
| National total estimated from the NFPA survey | fires, deaths, injuries or dollars | |
| Total of the same category reported to NFIRS | same as above | |
| NFIRS count for a specific problem within the category, such as one cause | same as above | |
| National estimate of that specific problem | same as above |
Separate ratios are used for residential structure fires, nonresidential structure fires, vehicle fires and other fires. Within each group, different ratios apply to fires, civilian deaths, civilian injuries and direct property damage.[7][10] Hall and Ahrens (2002) give an example for 1996: the survey estimated 4,080 civilian deaths in residential structure fires and NFIRS held 1,504, so the scaling ratio was 2.71.[10] As an illustration, an NFIRS subtotal of 100 residential structure fire deaths for one cause would then correspond to a national estimate of about 271 deaths. Because the survey is completed by summer each year, it produces numbers about a year earlier than NFIRS.[7]
Some fires never enter either database. Fires handled privately, by individuals or by private brigades, and fires reported only to wildland or military agencies are not captured.[7] A fire is also lost to NFIRS if no report is completed, if the report is not sent to the state, or if it contains a serious or fatal error.[7] Unusual events are hard to estimate. With North Carolina and Pennsylvania outside NFIRS in the 1990s, national estimates showed no dormitory or fraternity house deaths for 1993–1997, although FIDO knew of fatal fires in both states. Such projections can be footnoted, but not adjusted, on the basis of anecdotal data.[7]
Uses in fire protection
Fire protection engineers use fire experience data to support performance-based codes, define fire scenarios and estimate probabilities of fire and of protection performance. Fire services, planners, educators, insurers and product developers are among the other users.[22] In a fire hazard analysis, statistical data from NFPA or NFIRS on ignition sources, first items ignited and rooms of origin can show which factors contribute most to fires in the occupancy of interest.[26] Hall and Ahrens (2002) advise engineers in the United States to start from national estimates that use NFIRS for detail and the NFPA survey for projection.[10]
NFIRS counts incidents, not exposure, so it cannot by itself measure risk. Denominators such as prevalence in the population are needed.[27] A rate such as fire deaths per million population therefore divides an estimated number of deaths by a population count, as described under fire statistics.[28]
For fire departments, fire record analysis using NFIRS yields insights into hazards and trends that allow prevention to be targeted more effectively.[29] Hall (2008) calls national estimates from NFIRS and the NFPA survey the best national data for designing public fire education. Because NFIRS covered only about half of reported fires, analyses below national level were not possible for every area.[30]
A study of passenger vehicle fires shows how the databases are combined and checked. Using the NFPA survey, about 480 civilians died each year in highway vehicle fires in 1999–2004, and NFIRS proportions assigned about 88 percent of these, or 425 a year, to passenger vehicles.[31] This was about 8 percent higher than the 390 a year derived from the federal Fatality Analysis Reporting System. Long et al. (2008) concluded that NFIRS data appear to underestimate the share of fires and deaths associated with collisions, probably because other factors were recorded. Roughly half of passenger vehicle fires had no contributing factor recorded at all.[32] They also noted that the NFPA survey follows up fatalities to separate fire deaths from crash trauma, whereas NFIRS does not.[31]
Data quality and limitations
Quality control relies on standard edit checks that all Version 5.0 software must implement; incidents that fail are corrected by the state or returned to the department.[27] Software checks allowable values, but computers have no common sense: data that meet the criteria pass through, so a detailed narrative remains important.[33] For a database with the depth and breadth of NFIRS, the level of quality control possible for the NFPA survey, including follow-up calls, has not been possible. As a result, NFIRS has more missing and more dubious entries.[20]
Missing data are common, and severe fires tend to have larger shares of unknowns.[34] Reports on fatal and large-loss fires are often left incomplete pending investigation and never updated.[4] In nonconfined home structure fires in 1999–2002, the item first ignited was undetermined or not reported in 21 percent of fires and 39 percent of civilian deaths. It was missing for 18 percent of injuries and 30 percent of direct property damage.[35] Analysts usually allocate such unknowns in proportion to the known cases, which assumes that the unknown fires would look like the known ones. No allocation method had been shown to be the best, and analysts are expected to state and justify their scheme.[34]
Changes in the system can look like changes in the fire problem. The share of NFIRS structure fires coded as confined rose from 2.8 percent in 1999 to 34.2 percent in 2003, a large part of the rise being attributed to more departments reporting in Version 5.0.[34][36] Ahrens et al. (2008) advise checking whether any trend shift could be due to NFIRS changes rather than real fire experience.[8] A dislike of paperwork may also have led to undercounting of minor incidents, and there is no systematic way to control for underreporting or for changes in reporting practices.[16]
All fire department databases miss the fires that are never reported. Past studies cited by Schwartzman (2008) estimated unreported fires at 10 to 20 times the number of reported fires, although most were small cooking fires.[37] By their nature, fire databases are biased towards failures: a fire controlled so quickly that it need not be reported is not captured.[20] The Australian Building Codes Board's handbook for its Fire Safety Verification Method (2020) notes that design-fire frequencies are commonly derived from published fire incident data. It adds that many smouldering and small flaming fires are dealt with by occupants and go unreported.[38]
Comparable systems in other countries
Other countries organised fire data differently when the handbook was written, and some had systems related to NFIRS. Canada also had a three-tiered system, but its provinces reported totals rather than individual incidents to the national level.[5] The Australian Incident Reporting System, used by most Australian fire services, used a data classification standard based on NFIRS, with local modifications.[5] In Sweden the insurance industry collected fire data and only fire deaths were reported to the government. In the United Kingdom and Japan, fire services reported directly to a national government office.[5] Because definitions vary from country to country, Ahrens and Stewart (2008) advise caution in international comparisons.[5]
See also
Notes
- ↑ 1.0 1.1 1.2 1.3 1.4 Ahrens and Stewart (2008), p. 3-43.
- ↑ 2.0 2.1 Ahrens and Stewart (2008), pp. 3-43–3-44.
- ↑ 3.00 3.01 3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 Ahrens and Stewart (2008), p. 3-44.
- ↑ 4.0 4.1 4.2 4.3 4.4 4.5 Ahrens and Stewart (2008), p. 3-47.
- ↑ 5.0 5.1 5.2 5.3 5.4 5.5 Ahrens and Stewart (2008), p. 3-58.
- ↑ 6.0 6.1 Ahrens et al. (2008), p. 3-79.
- ↑ 7.0 7.1 7.2 7.3 7.4 7.5 7.6 Ahrens and Stewart (2008), pp. 3-55–3-56.
- ↑ 8.0 8.1 8.2 Ahrens et al. (2008), p. 3-61.
- ↑ Hall and Ahrens (2002), p. 5-69.
- ↑ 10.0 10.1 10.2 10.3 10.4 10.5 Hall and Ahrens (2002), p. 5-67.
- ↑ Custer and Sheppard (2008), pp. 3-31–3-32.
- ↑ 12.0 12.1 12.2 12.3 12.4 12.5 Ahrens and Stewart (2008), p. 3-48.
- ↑ 13.0 13.1 13.2 13.3 Ahrens and Stewart (2008), p. 3-49.
- ↑ 14.0 14.1 14.2 14.3 14.4 14.5 Ahrens and Stewart (2008), p. 3-50.
- ↑ Ahrens et al. (2008), pp. 3-70–3-71.
- ↑ 16.0 16.1 Ahrens et al. (2008), p. 3-65.
- ↑ 17.0 17.1 Ahrens and Stewart (2008), p. 3-53.
- ↑ Ahrens and Stewart (2008), p. 3-51.
- ↑ 19.0 19.1 Ahrens et al. (2008), p. 3-67.
- ↑ 20.0 20.1 20.2 Ahrens et al. (2008), p. 3-91.
- ↑ 21.0 21.1 Ahrens et al. (2008), p. 3-63.
- ↑ 22.0 22.1 22.2 Ahrens et al. (2008), p. 3-62.
- ↑ 23.0 23.1 Hall (2002c), p. 1-201.
- ↑ Hall and Cote (2008), p. 3-4.
- ↑ Hall and Ahrens (2002), p. 5-68.
- ↑ Hurley and Bukowski (2008), p. 3-122.
- ↑ 27.0 27.1 Ahrens and Stewart (2008), p. 3-55.
- ↑ Ahrens et al. (2008), pp. 3-79–3-80.
- ↑ Denlinger (2008), p. 12-71.
- ↑ Hall (2008), p. 5-117.
- ↑ 31.0 31.1 Long et al. (2008), p. 21-5.
- ↑ Long et al. (2008), pp. 21-5–21-6.
- ↑ Ahrens and Stewart (2008), p. 3-54.
- ↑ 34.0 34.1 34.2 Ahrens et al. (2008), p. 3-76.
- ↑ Ahrens et al. (2008), pp. 3-75–3-76.
- ↑ Ahrens et al. (2008), p. 3-77.
- ↑ Schwartzman (2008), p. 5-88.
- ↑ Australian Building Codes Board (2020b), pp. 61–62.
References
- Ahrens, M., Frazier, P., & Kelch, G. (2008). Use of fire incident data and statistics. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-61–3-92). National Fire Protection Association.
- Ahrens, M., & Stewart, S. (2008). Fire data collection and databases. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-43–3-59). National Fire Protection Association.
- Australian Building Codes Board. (2020b). Handbook: Fire Safety Verification Method (Print version 1.1).
- Custer, R. L. P., & Sheppard, D. T. (2008). Fire loss investigation. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-31–3-41). National Fire Protection Association.
- Denlinger, R. F. (2008). Managing fire-rescue departments. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 12-51–12-78). National Fire Protection Association.
- Hall, J. R., Jr. (2002c). Statistics. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 1-193–1-202). National Fire Protection Association; Society of Fire Protection Engineers.
- Hall, J. R., Jr. (2008). Using data for public education planning and decision making. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 5-103–5-118). National Fire Protection Association.
- Hall, J. R., Jr., & Ahrens, M. J. (2002). Data for engineering analysis. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 5-65–5-78). National Fire Protection Association; Society of Fire Protection Engineers.
- Hall, J. R., Jr., & Cote, A. E. (2008). An overview of the fire problem and fire protection. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-3–3-30). National Fire Protection Association.
- Hurley, M. J., & Bukowski, R. W. (2008). Fire hazard analysis techniques. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-121–3-134). National Fire Protection Association.
- Long, R. T., Jr., Colwell, J. D., Ray, R., Grossman, H. L., Thomas, B., & Strassburger, R. (2008). Passenger vehicle fires. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 21-3–21-14). National Fire Protection Association.
- Schwartzman, P. (2008). Juvenile firesetting. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 5-87–5-102). National Fire Protection Association.