HIPRC: Best Practices

Fire and Burn Injury Interventions

Multi-faceted Interventions

Background

A number of community based injury intervention campaigns have been conducted. The goal of the campaigns reviewed were quite broad and directed at more than one type of burn injury. Often these campaigns attempted to address a variety of other injury issues such as motor vehicle crashes, falls, and poisoning.

Review of multi-faceted burn interventions:

Author	Clark, 2000
Study design and target population	Time series in Maine 1961 to 1994.
Intervention	Burn education in schools; burn prevention programs around wood stove heating.
Outcomes	Annual burn mortality and hospitalization rates
Results	Burn mortality decreased from 5.1 per 100,000 in 1960-79 to 1.4 in 1993-96. Hospitalization rates decreased from 34.8 per 100,000 in 1973-76 to 10.6 in 1995-98.
Study quality and conclusions	Their control was the US as a whole which had much lower rates than the state of Main during this time and did not decline as fast. Unclear exactly what are the interventions.

Author	Shani, 1998
Study design and target population	Before/after design. Elemantary schools in Beer-Sheva Israel.
Intervention	Education in schools focused on (1) increase children's awareness of the problem, (2) provide information about injury risk factors, (3) strengthen one's sense of injury control, (4) develop and reinforce injury prevention practices and skills.
Outcomes	Incidence of burns requiring hospitalization.
Results	Rate decreased from 2.39 per 1000 in 1982 to 0.68 per 1000 in 1991.
Study quality and conclusions	No control groups.

Author	Ytterstad, 1998
Study design and target population	Non-randomized controlled community intervention in Harstad, Norway with Tronheim as the control city.
Intervention	Lowering of tap water temp, use of guard rails on stoves, community education.
Outcomes	Burns to children <5 requiring admission.
Results	51.5% decrease in burns in Harstad compared to 18.1% increase in burns in Tronheim from 1985-88 to 1993-96.
Study quality and conclusions	Data appear to show a consistent decrease in injuries.

Author	Maciak, 1998
Study design and target population	Before/after design in Detroit 1985-1996 aimed at preventing arson at Halloween.
Intervention	Multi-faceted program to decrease arson.
Outcomes	Reported fires.
Results	Decrease in the number of fires.
Study quality and conclusions	No controls.

Author	McConnell, 1996
Study design and target population	Time series. Low income residents of Memphis Housing Authority housing.
Intervention	35 minutes fire safety training program presented to 2,340 adults.
Outcomes	Fire incidents.
Results	At baseline, there was one fire per 895 renter months. After intervention, there was 1 fire per 1078 months among those exposed to the program. During the post-period, there was 1 fire per 780 renter months in the unexposed tenants.
Study quality and conclusions	Poorly controlled study.

Author	Clark, 1992
Study design and target population	Non-equivalent control group design. State of Maine, 1973-1988.
Intervention	Maine Burn Program (burn units established, safe installation and operation of wood stoves, legislation requiring smoke detectors, preventive education for school children).
Outcomes	Hospital admissions or deaths.
Results	Annual number of deaths per million persons declined from 41 (1973-1980) to 25 (1981-1988), change significant IRR=0.61;(p<0.001) compared to rest of nation. Annual hospital admissions per million persons dropped from 401 to 301 (IRR=0.75). Hospital mortality rate in burn units not significantly different from the regional standard burn unit, Massachusetts General Hospital.
Study quality and conclusions	Statewide burn program effective; prevention credited. Developed effective data collection and linkage systems. Controlled for secular trend by comparing to rest of the U.S.

Author	Guyer, 1989
Study design and target population	Non-equivalent control group 9 cities and towns in Massachusetts, with 5 similar cities serving as controls over a 22-month time period
Intervention	SCIPP studies used the Project Burn Prevention approach (community intervention campaign focusing broadly on many aspects of injury prevention).
Outcomes	Safety knowledge and practices Incidence of hospitalized injuries
Results	No change in burn injuries
Study quality and conclusions	High baseline rates of smoke detector use in population. Message too broadly targeted, aimed at 5 types of injury. Poor penetration: only 42% of households reached.. Recommend selecting interventions with strong evidence of efficacy.

Author	Gallagher, 1985
Study design and target population	Non-equivalent control group design Intervention carried out in 2 Massachusetts cities; city A (n=200) and city B (n=85). Families with children under 6 living in substandard housing (1981-82).
Intervention	HIPP (Home Injury Prevention Project), a 3- pronged intervention utilizing regulation, education, and technology: Enforcement of State Sanitary Code Safety counseling by housing inspector Distribution and installation of electrical outlet plugs and smoke detectors.
Outcomes	Change in number of code and non-code hazards. Sampled 82 households at follow up.
Results	Mean number of non-code household hazards reduced from 13.1 to 6.6 (p<.005); no difference in communities. 18% improvement in storage of flammable substances; less than 10% improvement in presence of smoke detectors or fire extinguishers; mean water temperature 11 degrees lower; electrical covers in place.
Study quality and conclusions	Technology items and regulation items showed greatest improvement. Minor improvement for educational items. Successful pilot project. Combine building inspection and injury prevention. Sample not random, too small to evaluate changes in injuries.

Author	MacKay, 1982
Study design and target population	Interrupted time series design. Population of the Greater Boston , MA. Comparison communities in Western MA Project Burn Prevention
Intervention	Educational program with 3 components: media campaign (TV, radio, newspapers) school-initiated intervention, and community initiated intervention. 8-month duration
Outcomes	Incidence and severity of injuries from scald, flame, electrical or contact burns, or smoke inhalation.
Results	No effect of the intervention overall. There was a moderate, temporary reduction in rate of burn injuries attributed to the community-based intervention but an increase in burn incidence observed for the media campaign. The school campaign showed no change.
Study quality and conclusions	A classic study. The lack of measurable results illustrates the need to narrowly target the intervention strategy rather than using a global approach of reducing all burn injuries. The authors felt that random variation in burn incidence was the most likely explanation for reduction in injuries.

Author	McLoughlin, 1982
Study design and target population	Interrupted time series design. Population of the Greater Boston , MA. Comparison communities in Western MA Project Burn Prevention
Intervention	Targeted flame and scald burns to high risk groups. Educational program with 3 components: media campaign (TV, radio, newspapers) school-initiated intervention, and community initiated intervention. 8-month duration
Outcomes	Change in knowledge, incidence, and severity of burn injuries.
Results	Significant increase in knowledge for children. Greatest increase for direct intervention methods.
Study quality and conclusions	Well-executed study, but it shows that change in knowledge does not lead to change in behavior.

Author	Kelly, 1987
Study design and target population	Randomized controlled trial Conducted at a primary care clinic at Yale-New Haven Hospital. Population composed of inner city children being seen for primary care visits during the first year of life.
Intervention	Use of a three-part, developmentally-oriented safety course with active parent participation and feedback. Control group received standard safety messages.
Outcomes	Hospitalizations and emergency visits for injuries; hazards in home evaluated by home visit.
Results	No difference between the groups was found for injuries. Accidents OR=0.78, 0.32-1.87; ER visits OR=1.6, 0.46-5.9); Hospitalization OR=1.0 0.01-79.9). Fewer hazards were found in the home for the intervention group as measured by a hazard score (2.4 vs. 3.0, p=0.02). Smoke detectors were absent in most of the dwellings (85% intervention, 89% control).
Study quality and conclusions	Minimal influence of safety education using a repetitive individualized approach. Home safety evaluators and chart reviewers were blinded to group status. 76% of the intervention group and 74% of the controls completed the study. Home visits were made to 84% of this group.

Summary of multi-faceted burn interventions

There appear to be some data to support the effectiveness of well done commun ity intervention programs in decreasing the rate of fires and of fire/burn related injuries and deaths. The data appear consistent that some community wide prevention efforts can make a difference. However, the strength of the evidence is weak. There a re no well done randomized controlled trials.

Recommendations on multi-faceted burn interventions

The injury campaigns that have demonstrated success have used a narrow specific message aimed at a well defined target. Multi-disciplinary approaches (vi sual and print media, counseling, campaign feedback) and involvement of a broad section of the population (lay organizations, government, medical groups, business) offer the best chances of success. Select injury interventions with strong evidence of effi cacy and target these interventions to a group at high risk.

Recommendations for future research

Use epidemiological data to pinpoint the population at greatest risk. Design the intervention with an evaluation component built in from the beginning of the project. Examine the elements of successful injury prevention campaigns and adapt them to the problem you wish to tackle. Conduct RCT's of community intervention programs.