Childhood Injury Prevention Interventions

Community Campaigns

Background

Public information and education campaigns have been used to address the child pedestrian injury problem. Some of these campaigns have been aimed at the child, others at parents, and still others at drivers. Some have been done in conjunction with environmental changes, such as the "Safe Community" campaigns.

The August, 2001 update includes results from 3 community environmental intervention campaigns. Two of these do not include pedestrian outcomes but they were included as examples of community-wide traffic safety interventions.


Review of community campaign interventions:

Author

Cross, 2000

Study design and target population

Ecologic mixed study.

Cohort of 1603 children followed from age 6 to 9 years.

Three metropolitan communities in Perth, Western Australia.

Intervention

Child Pedestrian Injury Prevention Project (CPIPP), educational intervention for students, parents & teachers and community. Also includes environmental improvements to road crossings. High, and moderate interventions are compared to usual practice.

Outcomes

Road crossing and playing behavior; self report validated with observation and personal interviews in a subset (n=80) of children.

Results

No difference in self reported knowledge.

Children in high and moderate intervention groups significantly more likely to cross the road with adult supervision (p=0.013) and play away from the road (p=0.000) than the comparison group.

Study quality and conclusions

Difficult to draw reliable conclusions about intervention effectiveness. Observed behavior should be used as outcome measure.

Majority of results based on self-report (questionnaire) responses. SR responses agreed with observations 50% of time while in person interview responses had 80% agreement with observations.



Author

Stevenson, 1999

Study design and target population

Ecologic mixed study.

Three metropolitan communities in Perth, Western Australia

1995-1997

Same project as Cross, 2000.

Intervention

Child Pedestrian Injury Prevention Project (CPIPP)

Group 1: Community/environmental road safety intervention and a school based road/pedestrian safety education program.

Environmental interventions included traffic calming, new pedestrian footpaths, pedestrian road signs & refuges.

Group 2: School based road/pedestrian safety education program.

Group 3: Control with no road safety intervention.

Outcomes

Volume and speed of traffic measured on local access & local distributor roads.

Index of community activity developed to quantify community.

Environmental interventions and local road safety initiatives.

Results

Significant reduction in traffic volume on local access roads in high intervention community (group 1).

No change in distributor road volume in any group.

Index of community/environ-mental activity in Group 1, 2.3 times greater and 5 times greater compared to Group 2 and control group.

Study quality and conclusions

Study strengths:

Authors believe educational & environmental changes with decrease child pedestrian injury.

3 cities similar in child pedestrian injury rates, SES and geographically separated to minimize contamination of the interventions.

No measurement of child pedestrian injuries.



Author

Durkin, 1999

Study design and target population

Ecological time trend study.

New York, New York.

Children <17 years of age.

Pre-intervention 1983-1988.

Post-intervention 1989-1995.

Intervention

Harlem Hospital Injury Prevention Program.

Model Safety City traffic education program delivered in elementary schools.

Construction of 25 new playgrounds and refurbishing of 18.

Enhanced summer recreation programs.

Outcomes

Hospitalized pedestrian injuries and deaths.

Northern Manhattan Injury Surveillance System database ICD-9 codes and E-codes used to identify subjects.

Results

Adjusted rate ratio for pedestrian injuries 0.55 (95%CI .38, .79) during the intervention period relative to the preceding period.

Injury decline occurred in targeted areas pedestrian (traffic, and bicycledecline in ). No injury nongroups. -targeted areas or age.

Poisson regression used to quantify and test for changes in monthly incidence controlling for variability in annual incidence.

Study quality and conclusions

45% decline in pedestrian injuries to school age children post-intervention.

Appropriate statistical analysis used to evaluate change.



Author

Ytterstad, 1995

Study design and target population

Before-after study

Harstad, Norway 1985-1993. Children under 10 were particularly targeted.

Intervention

Comprehensive community-based program to prevent injuries. Used variety of strategies, including education, environmental modification, and legislation.

Outcomes

Pedestrian injuries treated at the Harstad hospital.

Results

Risk of pedestrian injury to children 15 and under at the end of the intervention compared to baseline was RR=0.63 0.56-0.71). There was no significant decrease in pedestrian injuries in people over 15.

Study quality and conclusions

While the results appear to indicate a significant reduction in risk of pedestrian injuries, there was no control group (i.e., another city) to control for secular changes which, for example, might be due to less walking by young children.


Author

Malenfant and Van Houten, 1989

Study design and target population

Before-after design in three Canadian cities.

All ages; no specific data on child pedestrians

Intervention

Eight individual strategies used together to increase the number of drivers stopping for pedestrians at crosswalks.

Outcomes

Observations of motorists yielding to pedestrians at crosswalks.

Pedestrian-motor vehicle collisions.

Results

Percentage of motorists yielding to pedestrians in the 3 cities increased form 54%, 9% and 44% to 81%, 68% and 71% respectively. Pedestrian injuries decreased by 50%. (IRR = .50).

Study quality and conclusions

The intervention was well-designed and appeared to be successful in getting drivers to yield as well as decreasing the number of pedestrian injuries.


Author

Blomberg et al., 1983

Study design and target population

Before-after design, without comparison cities

Approximately 9000 children in Grades K-6 in Los Angeles, Columbus and Milwaukee, 1976-1977.

Intervention

Anti-dart out messages (Willy Whistle): 6 minute classroom film; poster; 30 and 60 second TV spots. Materials were distributed to schools and TV stations

Outcomes

(1) Street crossing behavior as children walked hone from school: looked left-right-left and full stop at the curb.

(2) Dart out injuries involving children 14 years and under.

Results

No effect on stopping at curb (RR=1.00, 0.99-1.02)

Modest effect on not looking (RR=0.94, 0.93-0.95)

Dart-out injuries decreased in target age group by 18% in LA, 36% in Columbus, and 18% in Milwaukee. Overall risk of dart-out injury after compared to before was RR=0.81, 0.78-0.84, adjusted for changes in numbers of other kinds of pedestrian injuries.

Study quality and conclusions

Approximately 1/3 of children in LA and 20% of children in Milwaukee saw school film; no reliable data for Columbus. TV spots played 200-380 times per city.

The intervention appeared to have very modest, if any, effects on behavior. Therefore, the decrease in injuries is difficult to understand.



Author

Thackray, 1982

Study design and target population

Time series analysis of injuries, 1972 to 1980.

Denver, Colorado 1977-1980.

Intervention

Comprehensive community campaign consisting of engineering, education, enforcement, and public information.

Outcomes

Police-reported pedestrian injuries.

Results

Pedestrian injuries to all ages decreased by 18.5% (OR =0.82) in Denver compared to Phoenix and San Diego in the same time period. Pedestrian injuries to children dropped by 40% (no controls).

Study quality and conclusions

Data are difficult to interpret, and the validity of the comparison cities is unknown.

Summary of community intervention campaigns:

Community intervention campaigns have taken a number of forms, and unfortunately few have been evaluated. The study by Blomberg19 is purely an educational program aimed at improving street crossing skills in young children. While the campaign showed sizable reductions in child pedestrian injury rates, the reductions appear unlikely to be due to the intervention. There was no effect on stopping at the curb and a very modest effect on no looking before crossing. Thus, it is hard to understand how the reduction of dart out injuries by 19% could be due to the campaign.

The other six programs were more concerted community-wide efforts that involved a combination of education, engineering, regulation and enforcement. The size of the reduction in pedestrian injuries was strikingly similar in the Denver, Norwegian and New York City experiments-about 40-45%. The Canadian study also found a decrease in pedestrian injuries but there was no specific data on child pedestrians.

The two reports from Western Australia evaluated community campaigns that included community involvement, environmental ch anges and educational campaigns. The intensive campaign was compared to a community safety education program and a control community with no specific road safety interventions. The intensive campaign appeared to be the most successful although there was not adequate documentation of child pedestrian behavior and changes in pedestrian injury rates were not measured.

A recent evaluation of the Harstad Safe Communities program used hospitalization and emergency department data from the Norwegian National I njury Sample Registry to compare Harstad to 3 control communities. Age adjusted injury incidence rates & standardized rate ratios calculated for all injuries indicated no significant long-term effects of the community-based injury prevention program. (Kopjar, 1999) This study did not evaluate changes in child pedestrian injuries, one of the targeted interventions.

Recommendations on community intervention campaigns:

The studies involving comprehensive campaigns lead us to recommend that these types of programs be introduced and evaluated elsewhere. The lack of the effect of a purely educational campaign leads us to recommend that such campaigns NOT be introduced.

There is a clear need for rigorous evaluations to be conducted of these interventions in o ther cities and countries, preferably in a randomized controlled fashion in which communities or neighborhoods are randomized and injuries to children are tracked as the outcome.