Table 1

Studies of community based prevention programs targeting all injuries in children aged 0–14 years

StudyLocation, community population/size, and demographicsTime frame for interventionType of controlType of interventionOutcomes: measures/methodResults
AIS, Abbreviated Injury Score; CI, confidence interval; MVA, motor vehicle accident; OR, odds ratio; RR, relative risk; SES, socioeconomic scale.
Studies designed with community controls
Schlesinger et al, 196613Rockland County, New York State, USA n = 4041 homes n = 4616 children based on 1962 censusJanuary 1962−June 1963Control community nearby: n = 4063 homes n = 4106 children based on 1962 census. Matched on family demographics, SES, type of housing, social interestsIntensive education program.
 Organization of neighbourhood discussion groups. Monthly newsletters mailed to families. Distribution of printed materials and brochuresMedically attended injuries for children aged under 7 years obtained directly from physicians, dentists, and hospitals via data collector visits every two weeks. Three month reporting to account for seasonal variation from baseline December 1960 one year before program and then to November 1963 six months after program ceasedIncidence rates of all injuries per 1000 children under 7 years: no differences found between study group and control group for age and sex adjusted injury rates during the three years
Guyer et al, 198914Nine cities in Massachussetts, USA. Total population 139810September 1980–June 1982Five control communities matched initially from 1970 census data on demographics, total population 146866. NB: by 1980, control communities did actually differ demographically with higher proportions of poorer and Hispanic households.
 This led to higher baseline injuries SES controlled for in analysisTargeted injury counseling for parents of young children. School and community burn prevention education. Household injury hazard identification and control.
 Community-wide promotion of poison control telephone information service.
 Public education concerning poison prevention. Promotion of child automobile restraint useInjuries in children aged 0–5 years requiring emergency department or hospital treatment or resulting in death.
 Injury cases recorded at 23 hospitals based on a surveillance system 1979–82Crude incidence rates (per 10000 per year) for all targeted injuries (MVA, burns, falls, poisoning): Intervention: pre, 262.80; post, 297.11 Control: pre, 477.42; post, 523.96 SES adjusted odds ratio for all injuries in pre-intervention period v post-intervention for intervention community v control is 0.99 (95% CI 0.83 to 1.19)
Davidson et al, 199415 Safe Kids/Healthy Neighbourhoods CoalitionCentral Harlem, Northern Manhattan, USA. Non-targeted: children aged <5 years (1990)
 Targeted: Children aged 5–16 years (1990) 39.5% live below poverty line Baseline injury rates twice as high as control communityFirst three years of intervention: 1989–91. Injury rates from 1983–88 (pre-intervention) with injury rates during intervention period. Comparing targeted v untargeted age groups and injuries types in intervention and control communityControl community: Washington Heights 30.7% live below poverty lineCoalition formed represented by at least 26 official and voluntary organisations.
 25 new playgrounds constructed and 18 existing playgrounds refurbished and fenced. Targeted pedestrian safety education to grade 3 students. Bicycle helmet giveaway programs. Supervised recreational programs with adult mentoring including a dance program, art studio, baseball league programs. Removal of existing summer programs from streets to playgroundsNorthern Manhattan Injury Surveillance System + death certificates. RR of severe injury and mortality due to all targeted injuries (assault, MVA, outdoor falls, guns) in children aged 5–16 yearsCentral Harlem. RR 0.56 (95% CI 0.45 to 0.71). Washington Heights RR 0.68 (95% CI 0.52 to 0.87)
Svanstrom et al, 199516 Lidkoping Accident Prevention ProgramLidkoping, Sweden (municipality in Skaraborg County).
 Population 35949.
 Agricultural and manufacturing county1984–91Comparable community controls: four bordering municipalities using the same hospital (population 42000) and whole of Skaraborg County (population 280000)Implementation of WHO Safe Communities model. This included:
 (1) Formation of interdisciplinary group represented by health professionals, teachers, munucipal administration, police, Red Cross, and residents.
 (2) Educational campaigns for general safety, bicycle safety, child safety seat use.
 (3) Training of sport coaches, day and child care staff, and parents in safety and first aid.
 (4) Environmental changes—for example, playgrounds, bicycle lanes, gymnasiums.Hospitalisation discharge data of injuries. Change in injury rates (% per year) for boys and girls in Lidkoping compared with control communitiesBoys: Lidkoping: decrease 2.4%, control: increase 0.6%, Skaraborg: decrease 1.0% Girls: Lidkoping: decrease 2.1%, control: increase 2.2%, Skaraborg: decrease 0.3%
Petridou et al, 199717Naxos, Greece. Population 14465; island1994Comparable island community: population 3802Open town meetings attended by health professionals, journalists, teachers, traffic police, women’s leagues, town residents. Educational seminars for parents.
 Workshops for teachers targeting safety at school. Interactive courses with primary and secondary schoolchildren. Removal of hazards from playgrounds and school yardsSelf reporting of injuries over 255 days from households: diary format.
 NB: break in recording over summer period when island population swells with touristsNo difference in injury rates.
 NB: data collected in sentinel community only, not whole community. 0–18 years: RR 0.85 (95% CI 0.68 to 1.06) for all accidents and RR 0.79 (95% CI 0.60 to 1.06) for home accidents
Coggan etal, 200018Waitakere, New Zealand. Population 155000.
 Large urban multicultural.
 Third of population under 20 yearsEvaluation over three year period 1995–97Control community: size 147000, matched on demographics, new housing development, road safety and crime safety workersImplementation of WHO Safe Community model (no further information provided)Hospitalizations for injury. Morbidity data 1989–1998Significant improvement in children (p<0.05)
Lindqvist et al, 200219Motala, Ostergotland, Sweden. Population data not providedProgram started in 1985.
 Pre-data collected 1983–84. Post-data collected 1989Comparable neighbouring community (similar age and gender structure)Implementation of WHO Safe Communities model. This included:
 (1) Mass media safety education.
 (2) Training of nurses to provide age adjusted safety information to parents at annual health visits.
 (3) Production of video demonstrating safety modifications in the home.
 (4) Environmental maintenance of playgrounds, schools, and day care centres.
 (5) Workshops for sport coaches and referees.
 (6) Implementation of “Safe way to school” program.
 (7) Safe cycling program with subsidised bicycle helmets.All injuries in children aged 0–15 years treated at health care units and local hospital. Comparing pre-intervention (1984–14) and post (1989) data. Odds ratios giving change in injury rates post-intervention v pre-interventionMinor injury (AIS 1) intervention: OR 0.89 (95% CI 0.80 to 0.99). Moderate (AIS 2) injury intervention: OR 0.49 (95% CI 0.41 to 0.57). Severe injury (AIS 3–6) intervention: OR 1.28 (95% CI 0.72 to 2.27). All injury intervention: OR 0.74 (95% CI 0.68 to 0.81). Control: OR 0.93 (95% CI 0.82 to 1.05)
Studies designed with the intervention community as an historical control
Jeffs et al, 199320Illawarra, NSW, Australia. Intervention: Shellharbour. Population 230000Baseline data collected 1987. Community information campaign 1990–91Historical control (although control communities are mentioned their purpose seems to be for knowledge/attitudes component)Implementation of WHO Safe Communities model. This included:
 (1) Formation of intersectoral taskforce represented by 13 agencies.
 (2) Community information campaign via mass media.
 (3) Backyard cleanups.
 (4) Bicycle safety campaigns.Injury surveillance in four hospitals for children 0–14 years17% decrease at A&E departments for children for injuries pre-intervention v post-intervention in Shellharbour. (20731/100 000 in 1987 to 17288/100000 in 1991) 14% decrease in severe injuries (653/100000 to 586/100000)
Tamburro et al, 200221Shelby County, Tennessee, USA1990–97Historical controlImplementation of SAFE KIDS coalition; examples of activities include:
 (1) Mass media/television campaigns.
 (2) School and holiday safety festivals.
 (3) Bicycle rodeos and support of state helmet legislation.
 (4) Smoke detector hand out program and education.
 (5) Baby walker campaign.
 (6) Child safety seat campaign and checks.
 (7) School pedestrian education.Children <9 years presenting for unintentional injury to a children’s medical center, emergency department or outpatient clinicsRate of severe targeted injuries decreased from 3.5 per 1000 during first two years to 2.0 per 1000 in year 7 (1996). (RR 0.77 95% CI 0.66 to 0.90) Non-targeted injuries increased from 1.4 to 2.5/1000