HIPRC: Best Practices

Fires and Burns

Interventions to Increase Smoke Detector Use

Background

The effectiveness of smoke detectors is well established. Smoke detector use in residential housing increased from about 5% in 1970 to 67% in 1982.^11,12 A number of factors probably contributed to increased use; specifically, price reduction, marketing campaigns, programs by local fire departments, building code modifications, and legislation. By 1983, 29 states required smoke detectors in all new residential construction and 22 states required retrofitting for some types of housing.¹³ Nevertheless, homes that need protection the most (low income dwellings) are least likely to have working smoke detectors. Strategies to increase smoke detector use (legislation, public education campaigns and give away programs) are reviewed below.

Review of studies evaluating programs to increase smoke detector use:

Author	DiGuiseppi, 2000
Study design and target population	Systematic review of controlled trials to promote smoke alarms.
Intervention	Promotion of SD 13 RCT 13 controlled.
Outcomes	Ownership of SD. Ownership of working SD.
Results	Counseling and education had modest effects on owning an alarm (OR=1.26, 95% CI 0.87, 1.81) or having a functioning SD (OR=1.19, 95% CI 0.85, 1.66). Counseling as part of primary child care had greater effects on ownership (OR=1.93, 95% CI 1.04, 3.58) and function (OR=1.72, 95% CI 0.78, 3.78).
Study quality and conclusions	Excellent Systematic Review.

Author	Mallonee, 2000
Study design and target population	Oklahoma project.
Intervention	See below.
Outcomes	Injury rates 72 months after interventions.
Results	Injury rates per 100,000 pop. Decreased by 81% in the target areas compared to 7% in the rest of the city. Injury rates per 100 residential fires decreased by 76% in target areas and increased by 12% in the rest of the city.
Study quality and conclusions

Author	Douglas, 1998
Study design and target population	Evaluation of different methods to distribute SD.
Intervention	Comparison of 4 methods of participant solicitation (Canvass door-to-door, flyers in public places, mailed flyers, flyers placed on doors).
Outcomes	Comparison of number of SD distributed to total number of homes.
Results	3564 SD distributed. % of homes receiving SD were 31% for canvassing, 3% for flyers in public places, and 6% for other two methods. Cost was lowest for canvassing.
Study quality and conclusions	Canvassing door-to-door was the most effective method to distribute SD.

Author	Shults et al, 1998
Study design and target population	Cross sectional.
Intervention	3 programs. Minnesota: Home visits with installation. North Carolina: Some installations, some SD giveaways. Oklahoma: see below.
Outcomes	Home visits to detect if SD were working 3 years after intervention.
Results	436 homes visited, 88% of homes had at least one SD. 64% has at least one working SD.
Study quality and conclusions	No control group.

Author	Mallonee et al., 1996
Study design and target population	Non-equivalent control group design The section of Oklahoma City with the highest rate of residential fires.
Intervention	Smoke alarm give-away program combined with education and publicity campaign. Surveillance data used to target area with highest risk.
Outcomes	Deaths and injuries from burns in residential fires; injuries per 100,000 population and injury rate per 100 fires in Oklahoma City; comparing the south central section to the rest of the city. Number of operational smoke detectors 4 years post intervention
Results	80% decrease in annualized injury rate per 100,000 population (from 15.3 to 3.1; incidence-density ratio 0.20 (0.10-0.40) compared to an 8% increase in the rest of the city (from 3.6 to 3.9; incidence density ratio 1.10 (0.70-1.7). Annual rate of fires per 1000 homes decreased 25% in target area and 18% in rest of city. Home inspection found 45% of smoke detectors functioning 4 years later.
Study quality and conclusions	Well-designed and executed study used surveillance data to target intervention area. Analyzed 3 years of burn incidence data prior to intervention and 4 years post intervention to stabilize rates. Temporal variation minimized by analyzing injuries per fire and per population. Suggests that a portion of the decrease in injuries is due to campaign education and publicity. Smoke detector give-away is an effective strategy to decrease fire-related injuries in high rate areas.

Author	Gorman et al., 1985
Study design and target population	Before/after design Baltimore, MD, city census tracts at high risk for fires
Intervention	Smoke detector give-away program. Free smoke detectors given to citizens who called to request them (n=3720).
Outcomes	Installation and proper functioning of smoke detector Risk status of population reached via random sample of home visits (n=231) and in-person interview. 82.5% response rate from random home visits.
Results	81% of homes inspected had installed and functioning smoke detectors. High-risk population reached. Linear correlation between the likelihood of receiving a smoke detector and the prior rate of fire injury and death (r=0.90, p<.001).
Study quality and conclusions	Successful smoke detector give-away program required active participation of a high risk population. Total program provided 110,000 smoke detectors to Baltimore households. (sold at cost and given away)

Author	McLoughlin et al., 1985
Study design and target population	Non equivalent control group design 2 affluent suburban counties: Montgomery County, Maryland (MC) and Fairfax county, Virginia (FC).
Intervention	Legislation-1978 required retrofitting of smoke detectors in all homes, (MC) vs. only new construction (FC).
Outcomes	Homes with working smoke detectors. Compliance with building code as judged by home visits
Results	Higher percentage of working detectors in intervention county (83% vs. 70%), OR=2.1 1.01-4.34. Similar compliance with code in MC (42%) vs. FC (44%). Only 6% of intervention homes were without detectors vs.16% for control community.
Study quality and conclusions	Evaluation in 1983. Unenforced law obeyed because it conforms to community values. Authors believe penalties attached to the law increased compliance. These are low risk communities. Study limited to single family homes which limits generalizability.

Author	Miller, 1982
Study design and target population	Non-equivalent control group design. Suburban pediatric practice in Pittsburgh, PA (n=240).
Intervention	Education, physician counseling with the option to purchase smoke detector at office visit (n=120). Control group received usual/routine care (n=120).
Outcomes	Correctly-installed and working smoke detector as determined by home visit.
Results	88.8% consented to home inspection. Moderate baseline use of smoke detectors (42.6% of intervention group vs. 53% of control group). 47.3% of intervention group purchased units vs. none of the controls. Intervention group less likely not to have operational smoke detector (OR=0.75, 0.41-1.4).
Study quality and conclusions	Rental units less likely to have smoke detectors than owned housing units (18% vs. 55%). Brief intervention which included offering smoke detectors for sale was successful in this middle class practice setting.

Author	Thomas, 1984
Study design and target population	Randomized controlled trial, single blinded. Members of HMO in Kansas City, Kansas who selected 90 min. "well baby" classes. Randomized to experimental (n=29) or control group (n=26); children under 1 year.
Intervention	Module on burn prevention including coupon for reduced price smoke alarm added to usual well baby class curriculum; home visit after 4-6 weeks for both groups
Outcomes	Scores on Fire Safety Knowledge Test Home tap water temperature Smoke alarm installation
Results	Mean score on Knowledge test 20.3 (experimental group) vs. 18.9 (control group), p=0.0001. 76% of experimental group had safe water temperature vs. 23% of controls; 66% of experimental group lowered temperature setting vs. none of the controls, p=0.01. No significant difference in smoke alarm installation.
Study quality and conclusions	Single education session effective. However, the study was small and had the following weaknesses: selection bias-volunteers for classes; population employed, married and in late 20’s; post-test measures only for water temperature and smoke detector installation.

Summary of interventions to increase smoke detector use

Legislation requiring installation of smoke detectors in new and existing housing¹¹ when combined with multi-factorial community campaigns and reduced price coupons seem to be best way to increase use.¹⁸ Smoke detector give away programs have proven successful when high risk areas are targeted.^14,18 The successful programs have had one specific goal, installation of working smoke detectors and used multi-faceted community campaigns. Multi-focus general injury prevention campaigns seem to have little effect (see section on multi-faceted interventions). Two barriers to effective use of smoke detectors are proper installation and maintenance and the reluctance of renters to purchase the devices.^7,16

Recommendations on interventions to increase smoke detector use

Legislation requiring smoke detectors in all housing units should be universal. Building codes should be amended to include installation and inspection of smoke detectors. Strict codes backed up by inspection can also address the issue of substandard heating and electrical equipment. Product modification, specifically child resistant cigarette lighters¹⁸ and self-extinguishing cigarettes⁶ are primary prevention strategies which would prevent many fires from starting. These technologies have been developed, tested and found feasible.¹⁹

Recommendations for future research

Reliable low price smoke detectors are available throughout the country. Improvements in current models should focus on developing a long lasting power source. Combination smoke and carbon monoxide detectors should be considered.