LESSONS LEARNED
EMERGENCY MEDICAL SERVICES FIRE SAFETY INTERVENTION
FROM AN
Ronald G. Pirrallo, MD, MHSA, Charles E. Cady, MD
The
public health approach to injury prevention is a systematic process that defines an injury problem, identifies risks and protective factors, develops and tests prevention interventions and strategies, and promotes widespread adoption of effective interventions and strategies.1 Slowly, traditional emergency medical services (EMS) systems have been adopting this paradigm.2 Prehospital care provided by dualtrained firefighter–emergency medical technicians represents a common EMS system model. This model is an ideal platform to combine the fire safety programs long championed by firefighters with access to an at-risk population served by EMS providers. As evidenced in the EMS Agenda for the Future, injury prevention should be a cornerstone of EMS.3 The risk factors for fire and burn death include smoking, young or advanced age, physical or mental illness, alcohol or drug use, and low socioeconomic or educational status.4–6 Not surprisingly, these populations are the highest users of emergency care as well. A 1998 pilot study by Pirrallo et al. found that persons living in one- and two-family dwellings who make multiple requests for fire department services, particularly EMS, are at an increased risk of experiencing a house fire.7 More importantly, only 17% of these dwellings had an operational smoke alarm at the time of the fire; nationally, the 1998 estimated average was 60%.8 Following the public health approach, having defined a problem (one- and two-family dwelling fire deaths), identifying an at-risk population (frequent users of EMS), and recognizing protective factors (operating smoke alarms), the Milwaukee Fire Department initiated a pubic health model fire safety intervention, ‘‘Operation First Responder.’’9 Our study objective was to assess the effectiveness of the Milwaukee Fire Department’s home fire safety intervention. We hypothesized that one- and two-family households that received smoke alarms and/or batteries during the Operation First Responder program were more likely to have an operational smoke alarm, less property dollar loss, and decreased morbidity and mortality at the time of a subsequent fire.
ABSTRACT Objective. The authors conducted a pilot study, finding that many households that experienced fires had received prior emergency medical services (EMS) visits, but few had operational smoke alarms. The study hypothesis is that dwellings that received smoke alarms and/or batteries during an EMS call were more likely to have an operational alarm, less property dollar loss, and decreased morbidity and mortality at the time of a subsequent fire. Methods. Smoke detectors and batteries were provided to an urban fire department for placement in unprotected homes at the time of an EMS call from March 1, 1999, through January 31, 2001. After addressing the reason for the 911 EMS call, verification or installation of an operational smoke alarm was performed. The authors examined records for dwellings that had a subsequent fire for outcomes of smoke alarm status, estimated property dollar loss, and number of injuries and fatalities. Results. This program placed 1,335 smoke detectors. Of these, 99 dwellings were found to have a fire or smoke condition with 20 exclusions. Our final number was 79; 28 (35%) still had an operating smoke alarm. In homes with operational alarms, the mean dollar loss was $2,870 (U.S. 2001) (95% confidence interval [CI], 143–5,596). In homes without operational alarms, mean loss was $10,468 (U.S. 2001) (95% CI, 5,875–15,061). No injuries or fatalities occurred in either group. Conclusion. This program was successful in placing 1,335 smoke alarms in at-risk dwellings and reaffirmed that an operational smoke alarm significantly decreases property dollar loss. However, if the goal is to have all homes protected by smoke alarms, this program has longterm effectiveness limitations. Key words: protection devices; accident prevention; fires; emergency medical services; smoke; burns.
PREHOSPITAL EMERGENCY CARE 2004;8:171–174
Received July 14, 2003, from the Department of Emergency Medicine, Medical College of Wisconsin (RGP, CEC), and the Wisconsin Injury Research Center (RGP), Medical Services, Milwaukee County EMS (RGP, CEC), Milwaukee, Wisconsin. Revision received and accepted for publication October 20, 2003. Presented in part as an oral abstract at the National Association of EMS Physicians Annual Meeting and Scientific Assembly, Tucson, Arizona, January 2002, and the 6th World Conference on Injury Prevention and Control, Montreal, Quebec, Canada, May 2002.
METHODS
Supported in part by the Allstate Foundation.
Study Design
Address correspondence and reprint requests to: Ronald G. Pirrallo, MD, MHSA, Froedtert Hospital East, 9200 West Wisconsin Avenue, Milwaukee, WI 53226. e-mail:
.
We performed a retrospective case series analysis with an intervention phase (placement of smoke alarms) from March 1, 1999, through January 31, 2001, and a
doi:10.1016/j.prehos.2003.12.009
171
172
PREHOSPITAL EMERGENCY CARE APRIL / JUNE 2004
six- to 28-month evaluation phase ending July 30, 2001. The Medical College of Wisconsin Human Research Review Committee awarded this study exempt status.
loss as estimated by the incident commander; and number of injuries and deaths.
Setting The Milwaukee Fire Department is the first responding agency to all emergency 911 calls serving a population of nearly 600,000, of which 55% are minority households.10 Milwaukee Fire Department personnel are dual-trained as firefighters and emergency medical technicians. Milwaukee Fire Department dispatching is based on a three-tiered system and responds to over 50,000 calls for service each year. All Milwaukee Fire Department first-responding fire engines and trucks were stocked with nine-volt zinc batteries and selfadhering, ionization sensor, no-pause button, zinc battery-operated smoke alarms. All personnel attended a 30-minute inservice on smoke alarm and battery installation and documentation requirements during scheduled refresher training.
Experimental Protocol When responding to a one- or two-family dwelling, and after addressing the reason for the 911 call, the company officer would, at his discretion, test existing smoke alarms and install an alarm and/or battery as needed to leave the home with an operational smoke alarm, the goal of the fire safety intervention. The tenant of the dwelling signed a consent form written in English and Spanish authorizing the installation and accepting responsibility for maintenance of the alarm. To enhance cooperation, policy suggested testing and installations be performed between 8:00 AM and 10:00 PM only. The consent form contained the following information: tenant’s name, address, and phone number; date of visit; number of smoke alarms present; number of operating smoke alarms; number of smoke alarms installed; number of batteries installed; and where the smoke alarms were located. Information from completed forms were entered into an Excel (Microsoft Corp., Bellevue, WA) spreadsheet database. For the evaluation phase, all alarms of fire or smoke condition in the city of Milwaukee from March 1, 1999, through July 31, 2001, were identified. The on-scene fire incident commander documents an alarm of fire or smoke condition using the National Fire Incident Reporting System (NFIRS; U.S. Fire Administration, Emmitsburg, MD). Milwaukee Fire Department administrative personnel transcribe their written report into the NFIRS electronic database. This database contained the following information: date and time of alarm; address; type of alarm; type of property; cause of fire; contributing factors; status of smoke alarm; whether the alarm alerted occupants; property dollar
VOLUME 8 / NUMBER 2
Analysis and Outcome Measures We verified that addresses in each database were in identical format. Selected address verification was made by performing a City of Milwaukee Assessor’s Office Property and Assessment Data Inquiry.11 The Assessor’s Office database contains property information, including type of dwelling (e.g., residential or business), size of unit (e.g., single, two, three, or multiple family), and street address. We eliminated incomplete or unverifiable addresses. The two databases were then cross-matched to identify dwellings that received the Operation First Response intervention and had a subsequent alarm of fire or smoke condition. We noted our primary outcome of finding an operational smoke alarm at the time of a subsequent alarm. We also recorded our secondary outcomes of property dollar loss, number of injuries, fatalities, cause of fire, and days between the initial EMS call and alarm of fire or smoke. We analyzed the outcomes with descriptive statistics: mean and 95% confidence interval (CI).
RESULTS During the study period, Milwaukee Fire Department personnel placed smoke alarms or batteries in 1,335 dwellings. Of these dwellings, 99 were found to have a subsequent alarm of fire or smoke condition. Twenty dwellings had to be excluded; 11 had incomplete records and nine incidents would not have benefited from a smoke alarm (e.g., vacant property or car fire). Our final number was 79. At the time of the subsequent alarm, 28 of these 79 (35%) still had an operating smoke alarm. The average number of days between the initial EMS call and subsequent alarm was 482; for homes with operating smoke alarms 399 days and without operational smoke alarms 369. The most common cause of fire was undetermined at the time of report completion. Anecdotally, no occupant refused inspection or installation of a smoke alarm. In homes with operational smoke alarms, the mean property dollar loss was $2,870 (U.S. 2001) (95% CI, 143–5,596). In homes without operational smoke alarms, mean loss was $10,468 (U.S. 2001) (95% CI, 5,875–15,061). During our study, no injuries or fatalities occurred in either group.
DISCUSSION All homes should have working smoke alarms. It is well known that smoke alarms, when installed and operating properly, save lives and reduce property
Pirrallo and Cady
173
EMS FIRE SAFETY INTERVENTION
damage and injury.8 Likewise, our study demonstrated a significant difference in estimated mean property dollar loss. Reducing residential fire injury and death remains a level 1 priority for the National Center for Injury Prevention and Control.1 To this end, several unique approaches have been attempted to improve household fire safety such as ‘‘change your clock, change your battery’’ campaigns and smoke alarm giveaways.12,13 Typically, these types of interventions have limitations and address only certain geographic areas of a community. A more effective and economic approach to home fire safety is to identify the populations that are at most risk for fire. The goal of Operation First Responder was to increase the home fire safety of an at-risk population. Although only 79 dwellings were identified as having an alarm of fire or smoke during our 28-month evaluation phase, the program ‘‘protected’’ over 1,300 dwellings. Disheartening and perhaps the basis of our lesson learned is that only 35% of the dwellings had an operational smoke alarm at the time of the fire. The greatest challenge in injury prevention is to implement what has been shown to be efficacious.14 The recently published randomized, controlled British study by DiGuiseppi et al. found a similar disappointing smoke alarm operation rate of 17% after a giveaway program.15 The even lower operational rate in this study could be explained because the majority of the smoke alarms were only given to the occupants of the household and few were actually installed. It appears that the power source and type of alarm contribute greatly to the likelihood of sustaining an operational home smoke alarm.8,16–18 Clearly, a passive prevention strategy of hard-wired, not batteryoperated, smoke alarms seems prudent. So, was Operation First Responder a failed public health approach to EMS-based fire safety? We did not demonstrate a lasting effective approach to improving residential fire safety, yet we learned several lessons. First, we demonstrated the importance of validating a pilot study. Our pilot study suggested that distributing and installing smoke alarms to this at-risk population would be a definitive solution; it is not. Second, fire safety is a complex multifactorial problem such that simply giving away smoke alarms will not improve residential fire safety greatly. Third and most interesting perhaps, we have uncovered a way of accessing an at-risk population that could be applicable to other injury control and prevention interventions. Areas of interest could include elder fall reduction, water heater temperature monitoring, and domestic violence screening. This plausible model best integrates EMS, public health, and injury control. The public health question remains, how can more at-risk homes be left with operational smoke alarms? In nearly two-
thirds of the dwellings, our intervention was not sustained. Apparently, factors other than the installation of a battery-operated smoke alarm need to be addressed to assure long-term alarm operability.
LIMITATIONS Our study had multiple limitations. First, our study population was a small convenience sample. This is the result of the fortunate fact that although identified as an at-risk population, the vast majority of addresses requesting EMS services did not go on to have a fire. Only 8%, or 4,238 of 52,468, of all Milwaukee Fire Department incidents were residential fires during the study period. Second, the testing and placement of smoke alarms were at the discretion of the commanding officer on the first responding unit. As a result of the need to address the acute emergency, not all homes with MFD visits were ensured of having an operational smoke alarm. Third, property dollar loss was based on the incident commander’s estimate, but cannot account for a greater than 3.5-fold difference between protected dwellings and those without operational smoke alarms. The incident commander was most likely blinded to a dwelling’s participation in Operation First Responder unless he or she responded to the initial EMS call and installed a smoke alarm or batteries. Fourth, we chose our surveillance period of this rolling enrollment study for convenience, because the ideal surveillance period for such a program is unknown. Fifth, it was impossible to confirm that the dwelling occupant remained the same during the surveillance period, and the contribution of the occupant to the resultant fire was unknown. Finally, we did not perform a cost–effectiveness analysis on this intervention. However, expenses were minimal because personnel were already on duty and smoke alarm and battery costs were approximately $6,000.
CONCLUSION Operation First Responder was successful in placing smoke alarms in 1,335 at-risk households, resulting in an operational smoke alarm at the time of a subsequent fire 35% of the time. None of the fires had any injuries or fatalities. Reaffirmed was the fact that the presence of an operational smoke alarm decreased property loss. However, if the goal is to have all homes protected by operational smoke alarms, this battery-powered smoke alarm giveaway program has long-term effectiveness limitations. The authors thank Milwaukee Fire Department Chief Lawrence Gardner (retired), Assistant Chief Mark Sain, Deputy Chief Gloria Murawsky, RN, EMT-P, and Captain Patrick Sierra, EMT-P, and recognize the men and women of the Milwaukee County EMS
174
PREHOSPITAL EMERGENCY CARE APRIL / JUNE 2004
System for their dedication toward improving the health and safety of their community.
8. Ahrens M. US Experience with Smoke Alarms and Other Fire Alarms. Quincy, MD: National Fire Protection Agency, September 2001. 9. Gardner LA. Milwaukee fire department: first responder guidelines ½internal memo #99-33. Milwaukee: Milwaukee Fire Department, 1999. 10. Milwaukee Journal Sentinel. JS Online: Minorities as % of Total Population. Available at: http://www.jsonline.com/news/ census2000/mar01/minorities. Accessed April 30, 2003. 11. City of Milwaukee Assessor’s Office. Property Assessment Inquiry. Available at: http://www.ci.mil.wi.us/citygov/assessor/ assessments.htm. Accessed April 7, 2003. 12. National Center for Injury Prevention and Control. Efforts to Increase Smoke Detector Use in US Households: An Inventory of Programs. Atlanta: Centers for Disease Control and Prevention, 1996. 13. Martin B. Excellence award profile. NY fire department delivers pizza and fire safety message. ICHIEFS On Scene. 2002;16:3. 14. Pless B. Smoke detectors and house fires [editorial]. BMJ. 2002; 325:979–80. 15. DiGuiseppi C, Roberts I, Wade A, et al. Incidence of fires and related injuries after giving out free smoke alarms: cluster randomised controlled trial. BMJ. 2002;325:995–7. 16. Rowland D, DiGuiseppi C, Roberts I, et al. Prevalence of working smoke alarms in local authority inner city housing: randomised controlled trial. BMJ. 2002;325:998–1001. 17. Fazzini TM, Perkins R, Grossman D. Ionization and photoelectric smoke alarms in rural Alaskan homes. West J Med. 2000;173: 89–92. 18. Shults RA, Sacks JJ, Briske LA, et al. Evaluation of three smoke detector promotion programs. Am J Prev Med. 1998;15:165–71.
References 1. National Center for injury Prevention and Control. CDC Injury Research Agenda. Atlanta: Centers for Disease Control and Prevention, 2002. 2. Garrison HG, Foltin G, for the Consensus Project Steering Committee. Consensus Statement on the Role of Emergency Medical Services in Primary Injury Prevention. National Association of EMS Physicians. Presented to the National Highway Traffic Safety Administration and the Maternal and Child Health Bureau, February 1996. 3. Delbridge TR, for the Steering Committee. Emergency Medical Services Agenda for the Future. US Department of Transportation, National Highway Traffic Safety Administration, August 1996. 4. Marshall SW, Runyan CW, Bangdiwala SI, Linzer MA, Sacks JJ, Butts JD. Fatal residential fires: who dies and who survives? JAMA. 1998;279:1633–7. 5. Istre GR, McCoy MA, Osborn L, Barnard JJ, Bolton A. Deaths and injuries from house fires. N Engl J Med. 2001;344: 1911–6. 6. Ballard JE, Koepsell TD, Rivara F. Association of smoking and alcohol drinking with residential fire injuries. Am J Epidemiol. 1992;135:26–34. 7. Pirrallo RG, Rubin JM, Murawsky GA. The potential benefit of a home fire safety intervention during emergency medical services calls. Acad Emerg Med. 1998;5:220–4.
VOLUME 8 / NUMBER 2