Intended for healthcare professionals

Primary Care

Physical activity cost of the school run: impact on schoolchildren of being driven to school (EarlyBird 22)

BMJ 2004; 329 doi: (Published 07 October 2004) Cite this as: BMJ 2004;329:832
  1. Brad Metcalf, statistician1,
  2. Linda Voss, senior research fellow1,
  3. Alison Jeffery, research nurse1,
  4. Jenny Perkins, nurse technician1,
  5. Terry Wilkin (T.Wilkin{at}, professor of endocrinology and metabolism1
  1. 1 University Medicine, Peninsula Medical School, Derriford Hospital, Plymouth PL6 8DH
  1. Correspondence to: T J Wilkin
  • Accepted 24 May 2004


Physical activity is perceived as important to children's health, and concern has been expressed at the increasing use of motorised transport to school—the “school run.”1 We have measured the activity cost of the school run in young children in the EarlyBird study.2

Participants, methods, and results

We analysed data from 154 boys and 121 girls in their first year at 53 urban primary schools. The children wore uniaxial accelerometers (Manufacturing Technology, Florida) during waking hours for five consecutive schooldays and the weekend to measure physical activity.3 We considered activity during the journey to and from school (8 to 9 am and 3 to 4 pm weekdays), school time (9 am to 3 pm weekdays), non-school time (before 8 am and after 4 pm weekdays), the total school week (weekdays), the weekend, and the total week. We measured height, weight, and body fat as the mean of five skinfolds. We found mode of transport and school journey time by questionnaire and distance to school with the RAC's online route planner. We used the National Statistics Socio-economic Classification to find socioeconomic status. Our study had 80% power to detect a significant difference (P < 0.05) of at least 12% in activity during the journey to or from school and at least 8% in total weekly activity.

Twice as many children walked to school (185/275; 67%) as were driven by car, with no significant gender bias (97/154 (63%) boys; 88/121 (73%) girls; Pearson χ2 test P = 0.12). The median time taken to walk to school was 6 (interquartile range 5 to 10) minutes and the median distance was 0.7 (0.4 to 1.2) km with 155/185 (84%) children walking less than a mile (1.6 km), comparable with the national figure of 82%.4

What is already known on this topic

The proportion of children travelling to school by car in the United Kingdom almost doubled from 16% in 1986 to 30% in 1998

There is a popular perception that being driven to school compromises physical activity

What this study adds

Being driven to school does not affect the overall physical activity of 5 year olds

Mean activity recorded during the 10 journeys to and from school was significantly higher during that period (0.75 units or 18% higher; P < 0.001) among those who walked than those who travelled by car (table). However, total weekly activity was identical (difference 0.04 units or 0.1%; P = 0.97). Crucially, the additional activity recorded by walkers during the school journey was only 2% (0.75/37.6) of the children's total weekly activity.

Physical activity during a week comparing children who walked to school with those who travelled by car recorded by 275 children according to mode of transport. Values are mean accelerometer count (×105) (standard deviation) unless stated otherwise

View this table:

Although the proportion of walkers was highest in the lowest socioeconomic group (C 65/82 (78%); B 49/78 (63%); A 71/115 (62%); P = 0.02), the pattern of results was unchanged when we analysed each social group in turn (data not given). Analysis of only moderate and high intensity activity gave consistent results. The two groups did not differ significantly in either body mass index (walk 16.1 kg/m2 v car 16.2 kg/m2; P = 0.88) or sum of skinfolds (walk 4.10 cm v car 3.93 cm; P = 0.28).


Although children who walk to and from school record more activity in the process, the difference has no impact on total weekly activity. Those driven by car matched those who walked to school in overall activity levels. These results apply to 5 year olds, but the EarlyBird study is longitudinal and ideally designed to compare findings once the children move from primary school to secondary school.

Another study to have considered the impact of walking to school on physical activity in children reported that boys (though not girls) who walk to school are more active after school.5 However, measurement of after school activity included the walk home, making the analysis difficult to interpret. Curiously, girls who walked recorded no more physical activity in so doing than those who were driven. The patterns of activity in the present study were systematically the same for girls and boys, lending further robustness to our findings.

Whether children walk to and from primary school makes no difference to their total activity. This does not justify the adverse publicity given to the school run nor the government's perception of the school run's impact. There may be other benefits from walking children to their neighbourhood school, but physical activity does not appear to be one of them.


  • This article was posted on on 18 August 2004:

  • Contributors TW conceived the study, supplied the hypothesis, edited the manuscript and is guarantor. LV conceived the study and edited the manuscript. BM performed the analysis and wrote the manuscript. AJ and JP recruited the subjects and collected all data.

  • Funding Diabetes UK, Smith's Charity, S&SW NHS Executive R&D, Child Growth Foundation, Beatrice Laing Foundation, Abbott, Astra-Zeneca, GSK, Ipsen, Unilever, Diabetes Foundation, EarlyBird Diabetes Trust.

  • Competing interests None declared.

  • Ethical approval Plymouth local research ethics committee of the South and West Devon Health Authority (1999).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract