Twenty-two studies were included: 3 randomised controlled trials (RCTs), 7 non-randomised controlled prospective studies, 11 uncontrolled prospective studies and 1 controlled retrospective study. The studies recruited individuals and households.
The studies scored from 3 to 9 out of a possible 10 points on validity.
Targeted behaviour change programmes: 6 studies of 4 interventions, including 3 controlled studies, were identified. One RCT showed that the Walk In to Work Walk Out self-help package significantly increased the time spent walking to work (increase compared with control 1.93, 95% CI: 1.06, 3.52). Three prospective controlled studies found that the TravelSmart programme (leaflets, timetables, maps and free trial bus tickets given to households) shifted trips from cars by 5.5% (P<0.01), 3.6% and 4.4% of all trips. One uncontrolled study found a positive shift of 25% of all weekday trips after 11 months using a Bikebusters programme, while the other uncontrolled study found no clear positive shift with the Travel Blending programme.
Agents of change and publicity campaigns: 4 studies, including 2 controlled studies, were identified. One controlled study found no effect from a school travel coordinator. The other controlled study found a significant decrease (P<0.05) in cycling in the intervention area. Two uncontrolled studies found little or no evidence for a community publicity programme or a workplace programme.
Engineering measures: 6 studies, including 1 controlled study, were identified. The controlled study found a positive shift of 3% to cycling after improving and extending cycle routes, compared with the control. Four of the 5 uncontrolled studies found either no clear evidence of a positive shift or a negative shift with the intervention; the other study found a positive effect.
Financial incentives: 2 studies, including 1 controlled study, were identified. The controlled study found that offering costs of parking to workers resulted in a positive shift of 1% of commuting trips after 1 to 3 years (P<0.01 compared with control). The uncontrolled study found a negative shift using a toll ring (2.6% of all trips).
Provision of alternative services: 3 studies, including 2 controlled studies, were identified. One controlled study found no effect from a City CarShare club. One uncontrolled study found a positive shift (5% of all trips, P<0.001). One retrospective controlled study found no effect for a telecommuting centre.
Effects on health: 6 studies were identified. Two RCTs of targeted behaviour change programmes showed a significant increase in mental health, vitality and general health status (1 RCT) and maximum aerobic power, maximal treadmill time and lactic acid after 10 weeks (1 RCT).
Social distribution of effects: 11 studies were identified. There were insufficient data to adequately assess the social distribution of effects.