Table 1

Main findings of systematic reviews on health promotion interventions to improve health through transport

Modes of intervention	Quality Indices	Main results
RR, risk ratio; RCT, randomised controlled trial; NS, not significant at p<0.05 level.
Primary care based counselling to prevent childhood injury	4²³	Injury prevention counselling as part of routine health supervision increased car seat and seat belt use, decreased motor vehicle occupant injuries and decreased hospital visits for traffic injuries.
	7²⁴	School based and public/parent education to use bicycle helmets reduced hospital inpatient rates for bicycle injuries by up to 0.2% more than control group. Reductions in hospital admissions as a result of general injury prevention approaches showed 20% decrease in 1 study, but NS effects in other programmes.
Promotion of childhood rear car seats	6²⁵	The evidence is weak that either educational campaigns or legislation to encourage front and rear seat belt use and placing children in rear seats are effective in changing behaviour. At some ages, there was a decrease in placing children in rear seats or in using rear seat belts. A number of included studies did not show statistically significant effects of the intervention.
Health promotion and community based approaches to reduce unintentional injury	5^26– ²⁸	(<15 years old): Road environment modifications reduced accidents by 7–32%; package of engineering measures reduced accidental injuries by 25%; road safety education can reduce casualties from children emerging from behind a vehicle by 20%; cycle helmets associated with 48% and 70% reduction in hospital admissions and death, plus 23% and 28% reduction in non-head injuries over 2 year study period; child restraints and seatbelts reduced injury severity.
	7^29, ³⁰	(15–25 years old): Bicycle and motorcycle helmets reduced head and other injuries and motorcycle helmet legislation was followed by a 30% reduction in fatalities, its repeal by an increase of 25–40% (the effect of reductions in cycling and motorcycling rates in the population is unclear); raising the minimum drinking age above 18 is associated with decrease in young driver and passenger fatalities.
		No proven effect of: training in reducing motorcycle injury; enhanced driver education courses; school-based programmes, rehabilitation for drink drivers, and education on the effects of catastrophic injury.
		Programmes that unintentionally enable adolescents to drive at a younger age than they would otherwise may have a negative effect.
Driver improvement and education programmes	3³¹	24/59 included programmes resulted in statistically significant reductions in violations (4–21%) but 3/59 resulted in significant increase in violations of 9, 14 and 40%. Crash reductions of 6–32% in 10/59 included programmes but 3/59 resulted in crash increases of 20, 30 and 46%. No proven effect of individual vs group interventions, direct vs indirect approaches or targeting certain types of violation.
	7³²	RCTs show increase in crash involvement and violations as a result of high-school aged driver education courses. Ecological studies show both increases and decreases in crash involvement after driving education programmes and increases in licensure rates in 16–17 year olds.
Road safety campaigns	3³³	Average campaign effect for all campaigns is 7.6% improvement. Persuasive rather than educative approaches are more effective. Legislation alone is not effective but requires enforcement plus publicity. Prior qualitative research, emotional vs rational appeal, theoretical model basis vs none, and specific behaviour request, increase the effectiveness of campaigns. Prevalence of baseline knowledge is inversely related to potential for impact of campaign.
	4³⁴	All road safety campaigns show 7.0% reduction in accidents over and above the background temporal reduction in accident rates. Financial rewards are most effective, followed by enforcement + legislation combinations and in cities rather than rural settings.
Safety belt incentives	7³⁵	Campaigns that use tangible incentives (such as money, prizes and vouchers) lead to substantial short-term increases in safety belt use (mean effect 12.0% increase above baseline) but have more modest longer term effects (mean effect 9.6% increase above baseline).
		Campaigns were most effective in elementary schools, where incentives were given immediately rather than delayed, and where the initial baseline use of seatbelts was low.
	6³⁶	Educational campaigns: 1 found 5% increase in children in rear seats (p<0.05); 1 pilot programme found 30% increase in child restraint use in rear seats (p<0.05) in elementary schools, but other settings and placing children in rear seat were NS.
		Legislation requiring restraints when children were in front car seats had effects on the use of rear seats: 1 study found 19% increase; 1 study found 9% increase in <1 year olds, 2% in 1–4 year olds and decrease of 4% in 5–9 year olds but NS effects in 10–14 year olds; 2 studies found NS effects.
	7³⁷	Child restraint use in rear seats: 3 studies found increases of 11–16% (p<0.05); 1 study found decrease in restraint use of 10% in 1–4 year olds and 3% in 5–9 year olds but increase in <1 year olds (all p<0.05).
		Community and clinical programmes to increase <5 year olds’ car seat and seatbelt use have moderate but only short term effects. 3 RCTs showed 36% increase in car seat or seatbelt use.
Remediation of drinking and driving offenders	7³⁸	RCTs of probation and rehabilitation to reduce alcohol consumption and injury-related sequelae showed improvements in motor vehicle crash risks (RR 0.76–0.90) and injuries (RR 0.47 and 0.58) but probation and rehabilitation together may increase risk of injury (RR 1.06 NS).
	3³⁹	Programmes to treat drink drivers show non-alcohol related crashes were worse as a result of the intervention (mean 11% increase) but a small decrease in alcohol related crashes occurred (mean 7% reduction). More severe licence sanctions reduced crash rates by 1–7% but lighter sanctions increased crash rates by 7%.