Intended for healthcare professionals


Influence of changing travel patterns on child death rates from injury: trend analysis

BMJ 1997; 314 doi: (Published 08 March 1997) Cite this as: BMJ 1997;314:710

This article has a correction. Please see:

  1. Carolyn DiGuiseppi, senior clinical research fellowa,
  2. Ian Roberts, senior lecturera,
  3. Leah Li, medical statisticiana
  1. Child Health Monitoring Unit, Department of Epidemiology and Biostatistics, Institute of Child Health, University of London, London WC1N 1EH
  1. Correspondence to: Dr Diguiseppi
  • Accepted 6 December 1996


Objectives: To examine trends in child mortality from unintentional injury between 1985 and 1992 and to find how changes in modes of travel contributed to these trends.

Design: Poisson regression modelling using data from death certificates, censuses, and national travel surveys.

Setting: England and Wales.

Subjects: Resident children aged 0-14.

Main outcome measures: Deaths from unintentional injury and poisoning.

Results: Child deaths from injury declined by 34% (95% confidence interval 28% to 40%) per 100 000 population between 1985 and 1992. Substantial decreases in each of the leading causes of death from injury contributed to this overall decline. On average, children walked and cycled less distance and travelled substantially more miles by car in 1992 compared with 1985. Deaths from road traffic accidents declined for pedestrians by 24% per mile walked and for cyclists by 20% per mile cycled, substantially less than the declines per 100 000 population of 37% and 38% respectively. In contrast, deaths of occupants of motor vehicles declined by 42% per mile travelled by car compared with a 21% decline per 100 000 population.

Conclusions: If trends in child mortality from injury continue the government's target to reduce the rate by 33% by the year 2005 will be achieved. a substantial proportion of the decline in pedestrian traffic and pedal cycling deaths, however, seems to have been achieved at the expense of children's walking and cycling activities. changes in travel patterns may exact a considerable price in terms of future health problems.

Key messages

  • Injuries are the leading cause of death in children over 1 year old, and we assessed the extent to which trends in mortality from transport related injury reflected changes in children's travel patterns

  • Child mortality from accidental injury declined by 34% between 1985 and 1992, while children walked and cycled less distance and travelled substantially more by car in 1992 compared with 1985

  • Substantial decreases in deaths from road traffic accidents for pedestrians and cyclists were at the expense of walking and cycling activities

  • Car travel became safer for children, but the effect on mortality was largely nullified by large increases in the distances children travel by car

  • Although these changes are in accordance with government targets to reduce child mortality from accidental injury, the associated decline in children's physical activity may lead to future health problems


In 1990 injuries were the leading cause of child death after the first year of life.1 the national strategy for improving health, health of the nation, has set a target of a 33% reduction in child mortality from unintentional injury by the year 2005.2 focusing solely on reducing deaths from injury, however, may lead to unintended consequences for other aspects of child health. we used data from death certificates and national travel surveys to assess the extent to which trends in mortality from transport related injury reflect changes in children's travel patterns.


We obtained an anonymised data file of all deaths from injury in children (aged 0-14 years) between 1985 and 1992 in England and Wales from the Office of National Statistics. Each record included age, sex, external cause of injury (E code),3 and year of death. We excluded 28 records (0.5%) missing the specific age at death. Total deaths from accidental injury and subgroups for specific causes were defined by the E codes as shown in table 1). We used annual numbers of deaths rather than annual numbers of death registrations as published in routine mortality data. Death registrations significantly misclassify year of death, since up to a third of deaths from injury each year are registered in later years, often because results of inquests are pending.

Table 1

Number of deaths from unintentional injury in children aged 0-14 years during 1985-92 in England and Wales and cumulative declines in death rates between 1985 and 1992

View this table:

The national travel surveys periodically collect data on the average annual number of miles that British residents travel by various modes, based on seven-day personal travel diaries.4 We analysed unpublished data from 1985-6, 1989-91, and 1992-4 (Department of Transport, 1996) for children aged 0-14 years and resident in England or Wales. We interpolated the annual distance travelled by each mode for each year from 1985 to 1992 to calculate annual death rates of pedestrians, cyclists, and occupants of motor vehicles per mile travelled by walking, cycling, and riding in cars respectively.

We used published census data to calculate death rates from injury per 100 000 population. we found no changes in age distribution over time and so have presented crude rates. we quantified trends in deaths per capita and per mile for 1985-92 using poisson regression modelling, with death rate as the dependent variable and year as the explanatory variable.5 for specific causes of death, we recalculated death rates for the subgroups with fewer than 100 deaths annually using three-year moving averages (two-year averages for 1985 and 1992) to minimise effects of random variability. a detailed tabulation of deaths and death rates among children and teenagers in england and wales in 1992 will be published elsewhere.6


There were 5392 child deaths from unintentional injury in England and Wales between 1985 and 1992. Pedestrian deaths in road traffic accidents accounted for the largest portion of these deaths (table 1)). Other leading causes of death included fire and flames, motor vehicle traffic accidents, and suffocation and aspiration. The total mortality from unintentional injury declined 5.7% per year between 1985 and 1992, a cumulative decline of 34% (table 1)). Death rates also declined for each of the leading causes of death from unintentional injury (fig 1).

Fig 1
Fig 1

Changes in mortality from specific causes for children aged 0-14 years between 1985 and 1992

Boys accounted for 3558 deaths (66%) and girls for 1834 deaths (34%) (table 1)). The ratio of boys to girls varied considerably by cause of death, ranging from 1:1 for occupants of motor vehicles in road traffic accidents to more than 5:1 for cyclists in road traffic accidents. For several categories, cumulative declines in death rates seemed to differ by sex, although the precision of these point estimates was low, especially for girls.

Between 1985 and 1992, the average distance walked in a year by a child aged 0-14 declined by 20%, from 247 miles (398 km) to 197 miles (319 km). The average distance cycled in a year fell 26%, from 38 miles (61 km) to 28 miles (45 km). In contrast, the average distance that children travelled by car in a year increased 40%, from 2259 miles (3635 km) to 3158 miles (5082 km). Girls showed substantially larger declines in walking (30%) and cycling (64%) and smaller increases in car travel (35%) than did boys (18%, 30%, and 45%, respectively). Children aged 10-14 showed the largest decline in walking (35%) compared with children aged 0-4 (10%) and 5-9 (20%). Declines in cycling mileage did not vary by age.

Table 2) shows the cumulative declines in deaths of pedestrians, cyclists, and occupants of motor vehicles per mile travelled by foot, cycle, and car respectively along with declines per 100 000 population. because of small numbers in some categories, age specific and sex specific rates must be interpreted with caution. number of deaths and annual distance cycled among pedal cyclists aged 0-4 were insufficient for reliable trend estimates.

Table 2

Cumulative declines in deaths of children aged 0-14 years in road traffic accidents between 1985 and 1992 in England and Wales

View this table:


If current downward trends in child death rates from injury continue, we will have no difficulty achieving the national target set in the Health of the Nation.2 For transport related deaths, however, these declines seem to have occurred at the expense of children's walking and cycling activities. It is possible that the safety of walking and cycling has improved, as suggested by declines indeaths per mile walked and cycled. However, concurrent reductions in case fatality rates probably account for much of these declines.7 The much greater declines in pedestrian and cycling deaths per capita compared with deaths per mile reflect large declines in the distances children walk and cycle.

on the other hand, deaths of occupants of motor vehicles per mile travelled declined by almost half between 1985 and 1992, suggesting that, along with reductions in case fatality rates, car travel became a great deal safer. there was only a small effect of increasingly safe car travel on deaths of occupants of motor vehicles per capita, however, because children now travel many more miles by car each year.

Differences in mortality by age and sex

In general, in each age group and sex, the relative differences between the decline in transport related deaths per mile and the decline per capita were similar to the overall declines. Greater declines in deaths per capita among girls may reflect their increasingly restricted travel—that is, their greater decline in walking and cycling and smaller increase in car travel compared with boys.

Larger declines in pedestrian deaths per mile walked among children aged under 10 compared with older children may reflect an increasing tendency for young children to be escorted by adults,8 making walking a safer activity for them. At least some of the differential decline by age and sex in cycling deaths per mile might be due to differential increases in the use of cycle helmets.9

Reasons for the large differential declines by age and sex in deaths of occupants of motor vehicles per mile are unclear. they might be explained if there were differential trends that affected safety of car travel—for example, differences in use and effectiveness of safety restraints or in type of car journey—or that affected case fatality rates after injury.

Shortcomings of study

Deficiencies in completeness or reliability of death certificates are unlikely to have systematically biased our results: all the key variables were complete, underreporting of injury as the cause of death is rare for young people,10 and there were no changes in assignment rules for cause of transport related death during the period studied. Because of small differences over time in the representation of London residents in the national travel surveys,4 we may have slightly overestimated the contribution of declines in distance cycled, and underestimated the contribution of declines in distance walked, to downward trends in death rates among cyclists and pedestrians. These differences would have had little effect on our overall results, however.

Implications for child health

While declines in children's walking and cycling may reduce traffic related deaths, they can otherwise adversely affect child health.11 12 Reduced cycling and walking have undoubtedly contributed to declines in overall physical activity, a cause of increasing obesity in British children,13 14 and potentially of increased obesity in adulthood with its associated risk of chronic disease and death.15 The curtailment of independent mobility may also have important adverse effects on children's mental, physical, and emotional development.11 It is likely that declines in walking and cycling were smaller among children of lower socioeconomic status because they have fewer alternatives for play or transportation.12 16 This differential exposure to traffic related risk may have contributed to divergent trends in death rates from injury by social class.17

Reduced walking and cycling, and increased car travel, may thus exact important societal costs by increasing future health problems and widening socioeconomic disparities in child death rates. The targets in the Health of the Nation are part of the national strategy to improve health. For child injury deaths, we risk hitting the target but completely missing the point.


Funding: CD is funded by Camden and Islington Health Authority; IR and LL are funded by the Sir Siegund Warburg's Voluntary Settlement Trust.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.