FOR IMMEDIATE RELEASE
KEY REPORT QUESTIONS RELIABLITY OF NFIRS DATA
IN IDENTIFYING FURNITURE FIRE LOSSES AND SOURCES
Inconsistent Reporting, Lack of Statistical Confidence Suggest NFIRS Data Not Accurate Enough For Policy Making
CHICAGO (July 19, 2016)—The Fire Prevention Alliance (FPA) is raising concerns about the reliability of summary information regarding upholstered furniture reported by the National Fire Protection Association (NFPA) based on data from the National Fire Incident Reporting System (NFIRS).
An NFPA “White Paper on Upholstered Furniture Flammability” report estimates risks and fire-related costs (deaths, injuries, and property losses) attributable to upholstered furniture as the primary or secondary ignition source of residential fires. The paper draws from NFIRS data.
But, FPA believes that the NFIRS data do not accurately reflect risks related to upholstered furniture as an ignition source because of incomplete field reporting, and extensivemissing information and “unknown” entries. Because of this, says FPA, the NFIRS data is not accurate enough to be used for setting policy or regulations relating to upholstered furniture and fire safety.
The validity of the NFIRS data is critical, because it is often used by the Consumer Product Safety Commission (CPSC) and other regulators to help determine fire safety policies. Misinterpretation could result in regulatory initiatives that are too strict or too lax.
Study Evaluates Data Accuracy
FPA recently commissioned a study by The Brattle Group, an economic consulting firm focused on answering complex economic, regulatory, and financial questions for corporations, law firms, and governments, to review the NFIRS data and NFPA approach in detail.
Experts from The Brattle Group were asked to evaluate two estimates:
1) The number of deaths attributable to fires where upholstered furniture was identified as the item of first ignition (smolder + small open flame + other ignition sources);
2) The number of deaths attributable to fires where upholstered furniture was designated as the principal item responsible for fire spread (primary fuel source as an item of secondary ignition).
For the analysis, The Brattle Group report focused on home structure fires.
“The Brattle Group Report points out some serious inconsistencies in the NFIRS data,” notes Robert Luedeka of FPA. “It shows how there may be wide variations in how the information could be interpreted.”
Significant Inconsistencies Identified
Using the NFIRS data, NFPA issued findings that there had been a 67 percent decrease in deaths where upholstered furniture was the item first ignited between 1980 and 2009. However, NFPA also found, using the NFIRS data, an additional 130 deaths may be attributed upholstered furniture as a secondary ignition source and the principle source for flame spread in the period from 2006-2010. The NFPA “white paper” would indicate a 27 percent increase over the CPSC Residential Fire Loss count based on first ignition alone.
This implies that the number of fire deaths due to upholstered furniture as a primary or secondary ignition source has been underreported in the past and the current risk may be higher than thought. For example, as deaths per million, a standard risk measure, CPSC data provide a risk factor of 1.36 due to furniture ignition-related deaths and NFPA interpretation raises the risk factor to 1.77 when deaths attributed to furniture as the primary source of fire spread are included.
Experts at The Brattle Group, in their analysis of the data, found that the NFPA interpretation of NFRIS data is based on statistically unreliable estimates and NFRIS data are subject to uncertainty, making them of limited value for policy-making decisions. Examination of the NFPA analysis methods and raw NFRIS reported data showed a high degree of uncertainty in the estimates that may not be widely recognized and this lack of statistical reliability is not reported in NFIRS, NFPA, or CPSC documents.
“A lack of statistical reliability indicates possible large variations in the data, sometimes creating a range of uncertainty that’s wider than the estimates themselves,” noted Dr. Charles Gibbons, an Associate at The Brattle Group. “In just one example, in 2012, there were an estimated 452 deaths attributed to upholstered furniture as an initial source. But the confidence interval for this figure was +/- 246 deaths. So the actual number of deaths could be as low as 206 or as high as 698. A range this large makes it impossible to rely on these data for formulating effective policy.”
The Brattle team attributed the gaps in data reliability to several factors. NFIRS reports include both missing data fields and unknown ignition sources because fire department reports are often incomplete (in 2012, 30 percent of first response ignitions were missing and two percent were unknown; for flame spread, 75 percent were missing and 10 percent were unknown).
Allocation techniques applied by NFPA to overcome these data gaps may not represent the “real world” in various regions on the country. Unless NFPA’s unverifiable assumptions about these fires are correct, the resulting summary information may not be accurate.
In addition, the authors of the Brattle report concluded that the accuracy of fire department reporting is questionable when forensic analysis of a fire has not been performed. It cannot be determined whether missing information and “unknown” entries are the result of the lack of forensic analysis, pending analysis results, or simply unanswered and incomplete NFIRS reports.
“There are just too many uncertainties to be able to accurately interpret the NFIRS data as currently reported,” Robert Luedeka of FPA said. “The Brattle Group experts offered some suggestions as to how to improve the reliability of the data, such as studying a subset of fire departments that consistently report to NFIRS or studying regions that have low levels of non-response, but these alternate methodologies would need to be evaluated.”
He added, “the key take-away is that fire loss summaries based on NFIRS data should not be used in policy-making considerations. Statistical reliability of the basic NFRIS data is weak and further interpretation, including allocating data to support missing and unknown entries, will not improve reliability.”
A complete copy of The Brattle Group report is available at
About The Fire Prevention Alliance
The Fire Prevention Alliance is a coalition of various industries dedicated to fire prevention and safety in respect to home furnishings products. FPA members include the American Home Furnishings Alliance, the National Council of Textiles Organizations, the National Cotton Council, the Polyurethane Foam Association, and the Upholstered Furniture Action Council. The alliance has partnered with agencies that help to serve FPA audiences on varying levels, including the U.S. Department of Health and Human Services, the U.S. Department of Housing and Urban Development, public school systems, rural churches, and FPA members.
# # #