• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jepicomhInstructions for authorsCurrent TOCJournal of Epidemiology and Community Health
J Epidemiol Community Health. Jan 2007; 61(1): 34–39.
PMCID: PMC2465585

The population effect of crime and neighbourhood on physical activity: an analysis of 15 461 adults

Abstract

Area‐based interventions offer the potential to increase physical activity for many sedentary people in countries such as the UK. Evidence on the effect of individual and area/neighbourhood influences on physical activity is in its infancy, and despite its value to policy makers a population focus is rarely used. Data from a population‐based health and lifestyle survey of adults in northwest England were used to analyse associations between individual and neighbourhood perceptions and physical activity. The population effect of eliminating a risk factor was expressed as a likely effect on population levels of physical activity. Of the 15 461 responders, 21 923 (27.1%) were physically active. Neighbourhood perceptions of leisure facilities were associated with physical activity, but no association was found for sense of belonging, public transport or shopping facilities. People who felt safe in their neighbourhood were more likely to be physically active, but no associations were found for vandalism, assaults, muggings or experience of crime. The number of physically active people would increase by 3290 if feelings of “unsafe” during the day were removed, and by 11 237 if feelings of “unsafe” during the night were removed. An additional 8342 people would be physically active if everyone believed that they were “very well placed for leisure facilities”. Feeling safe had the potential largest effect on population levels of physical activity. Strategies to increase physical activity in the population need to consider the wider determinants of health‐related behaviour, including fear of crime and safety.

In the UK, as many as two thirds of adults live sedentary lives,1,2 representing one of the least physically active nations of 15 European member states.3 Lack of regular physical activity is associated with marked preventable mortality and morbidity4 and is a public health priority. Although efforts to increase physical activity among individuals have had some small effect,5,6,7 modifying social, economic and environmental factors may be more successful at the population level.6 Indeed, evidence is emerging that contextual or area‐level factors, including transport systems, land use mix, population density and leisure opportunities, are related to population levels of physical activity.7,8,9,10 However, few studies have examined this in the UK8 despite sedentary behaviour being a major public health concern. Increasing our understanding of the relationship between physical activity, “who you are” and “where you live”,9 although challenging,6 is essential to inform the development of interventions to seriously reduce the number of people living mainly sedentary lives.

In the general population, regular physical activity is more likely among men, younger adults, people with other healthy lifestyle behaviours (eg, non‐smoking, greater intake of fruit and vegetables), those reporting good general health and no history of chronic disease.2 The aim of the current study was to examine neighbourhood influences on physical activity and to quantify this in terms of the population effect using population impact measures (PIMs).10 PIMs provide a population perspective by adding incidence information to traditional measures of risk, such as the population attribu

risk, thus, providing information on the actual numbers of people who are at risk from specific exposures in a particular population to assist local policy decisions.11

Methods

The study was based in two districts in northwest England, which is divided into 44 administrative electoral wards. Data from the 2001 national census calculated a resident population of 567 600 adults: 94% were white people and the population density was 1700 people per square kilometre.12 Methods for data collection have been described previously.2,13,14 In brief, data were collected using a postal self‐completion questionnaire as part of a population‐based health and lifestyle survey in 2001. The sampling frame was all resident adults on the general practitioner register and systematic sampling was used to select a 5% sample. The postal questionnaire was sent with a covering letter and a business pre‐paid return envelope. Non‐responders were sent a reminder postcard 10 days later. After another 10 days, persistent non‐responders were sent a reminder letter with another copy of the survey and a return envelope. The questionnaire included an introduction in Gujarati and Urdu, the main second languages spoken in the area, with information on the local health translation services. A favourable opinion was received from the local research ethics committees before starting the study.

The 50‐item questionnaire sought information on general and specific health, health behaviours and perceptions of neighbourhood. Question constructs were taken from previous national surveys.1,14 The questions specific to the neighbourhood asked respondents about the following: the extent that they felt they belonged to that area (strongly agree to strongly disagree); how well placed their home was for public transport, general shopping and leisure facilities (very well placed to badly placed); in their neighbourhood how much of a problem was vandalism; assaults and muggings; speeding traffic; and whether they had been the subject of personal crime in the past year. They were also asked whether they felt safe “out and about” in their neighbourhood during the day and during the night. Multiple deprivation was measured using the Townsend Index, which is constructed on four census variables (unemployment, overcrowding, non‐car ownership and non‐home ownership).15 Townsend Scores from 1142 census enumeration districts for the two electoral districts in the study were assigned using the participants' postal code.16

Physical activity was assessed using the Godin and Shephard instrument.17 This is valid for use in epidemiological studies and discriminates between adults participating in different amounts and types of physical activity. Participants were asked to record how many times in the past week they had engaged in light, moderate or vigorous activity for a session lasting at least 15 min. Examples of moderate physical activity included brisk walking, table tennis, easy cycling, golf, dancing and cleaning windows; vigorous activity included running, football, cardiovascular gym workouts and aerobics. In the current analysis, physically active was defined as participating in at least five sessions per week of moderate or vigorous physical activity, with each session lasting at least 15 min.17

Analysis

Individual associations with physical activity and neighbourhood factors were expressed as relative differences (prevalence rate ratios) using a modified Poisson regression approach.18 This involves fitting a generalised linear model to the data with a log link and a Poisson error term. The outcome variable in these models was being physically active, and the predictor variables were the health and lifestyle behaviours. The robust variance estimator was used to adjust for misspecification of the error term. The analyses controlled for the potential confounding effects of age, sex, ethnicity and deprivation. Data were analysed Stata V.8.2 (StataCorp, College Station, Texas, USA).

The population effect of eliminating a risk factor was calculated when the relative risk was statistically significant. The calculation excluded a time element, given the cross‐sectional nature of our data. Its formula is10:

PIN‐ER=n×Ip×PAR

where n is the population size; Ip is the incidence of sedentary behaviour (physical inactivity) in the whole population; PAR is the population attributable risk (Pe(RR−1)/1+Pe(RR−1)); Pe is proportion of the population who is physically inactive; RR is relative risk.

Calculations of the population attributable risk for variables with multiple strata were adjusted according to the methods of Hanley.19

Results

In June 2001, 70.1% of the sample returned a useable questionnaire (15 461/21 923). Their mean age was 49.8 (standard deviation (SD) 17.6) years, 45.2% (6986) were men and 95.5% (14 765) described themselves as Caucasians. The mean age of responders was 8.3 years more than that of non‐responders. No other information on non‐responders was available for comparisons. In all, 27.1% (4193/15 461) of responders defined themselves as being physically active. The mean age of physically active respondents was 10 years lesser than those not defined as being physically active (42.5 v 52.5 years, p = 0.001).

We found no differences in the proportion of men and women who were defined as physically active (27.6% v 26.7%), but those described as Caucasians compared with non‐Caucasians had a higher relative prevalence of physical activity (1.32, 95% confidence interval (CI) 1.16 to 1.52). For deprivation, a graded relationship was observed, with the prevalence of physical activity reducing across each of the deprivation quintiles (table 11).

Table thumbnail
Table 1 Prevalence of physical activity by baseline characteristics

Looking at neighbourhood factors, a graded relationship was observed between how well people thought their neighbourhood was for leisure facilities and the prevalence of being physically active (table 22).

Table thumbnail
Table 2 Association of physical activity with individual perceptions of neighbourhood facilities

We found no association between physical activity and sense of “belonging” to their neighbourhood, how well placed they believed their neighbourhood was for public transport and for general shopping (table 22).

People who felt unsafe out and about in their neighbourhood during the day (relative prevalence 0.70, 95% CI 0.59 to 0.82) and during the night (relative prevalence 0.82, 95% CI 0.78 to 0.88) were significantly less likely to be defined as physically active compared with those who felt safe during these times (table 33).

Table thumbnail
Table 3 Association of individual perceptions of crime and safety with physical activity

We observed no association for physical activity and people stating that vandalism, and assaults or muggings were a problem in their neighbourhood, also not among people who had or not been victims of personal crime during the past year. People who thought that there was some problem with speeding traffic in their neighbourhood were more likely to be physically active, but this was not consistent to this being a serious problem.

Table 44 shows the population effect of eliminating statistically significant risk factors for sedentary behaviour.

Table thumbnail
Table 4 Estimated population effect on physical activity from changing neighbourhood perceptions

The data suggest that the number of physically active people would increase by 3290 if feelings of being unsafe during the day were removed, and by 11 237 if feelings of being unsafe during the night were removed. An additional 8342 people would be physically active if everyone believed that they were “very well placed for leisure facilities”. In absolute terms, this would be expected to increase the current level of physical activity in the population by 0.6%, 2.0% and 1.5%, respectively (table 55).

Table thumbnail
Table 5 Number of people in the total population expected to become physically active if neighbourhood perceptions improved

Discussion

Our work represents one of the most comprehensive assessments of individual and contextual associations with physical activity among adults in the UK general population. We have previously confirmed low levels of physical activity among several adults, which decreased with advancing age and by socioeconomic deprivation.2 The focus of the current investigation was to examine the association of physical activity with contextual factors, based on the notion that both individual and contextual factors can influence physical activity. We found that individual perceptions of how well placed their neighbourhood was for leisure facilities were considerably associated with physical activity. The fact that this increased across each response category adds strength to this dose–response association. We also found that feeling safe in the neighbourhood during the day or during the night was positively associated with physical activity. Our approach of applying population effect measures suggested that the greatest increase in physical activity would be achieved in the population if everyone was made to feel safe during the night, with only a small effect if everyone was made to feel that their neighbourhood was well placed for leisure facilities. Therefore, if we are to increase population levels of physical activity, increasing feelings of safety seems to be a greater priority than improving perceptions regarding the provision of leisure facilities.

In our study, we failed to find any consistent association between physical activity and sense of belonging to the neighbourhood or perceptions about transport or shopping facilities, or problems in the neighbourhood from unsociable and criminal behaviours. Perhaps these did not differ sufficiently across the study setting to influence physical activity or among this population these factors may have had little effect on this behaviour.

The strengths of our study are its population focus, a large sample size with good response rates and data on a wide range of possible effects on physical activity. The survey included validated questions and reflected those used in national surveys and surveillance systems. We also adjusted for the potential confounding effects of area deprivation, using the participants' postcode linked to deprivation data at an enumeration level. Although this method has been found to be an effective method to examine the effect of individual deprivation on health,16 some misclassification may have taken place.

PIMs are a recently described addition to other measures of population effect, such as population attributable risk.10 PIMs add information on incidence to estimate the number of people in a total population who may benefit (or be at risk) from an intervention. As such, they provide a population perspective to inform local policy decisions.16,20 In the current study, this method has been used to estimate the effect of neighbourhood and neighbourhood perceptions on sedentary behaviour in adults.

Our study relied on self‐reported measures, which may be subject to measurement error, and our control for confounders was limited to the data originally collected. Simple methods for assessing physical activity have been found to reliably predict outcomes such as mortality,21 supporting their wide application in epidemiological studies. Response bias is a known problem in population studies and just <30% of those in the original sample did not return a useable questionnaire. A previous study found that non‐responders were less likely to be physically active compared with responders.22 Therefore, the true prevalence of sedentary behaviour in the population studied might be more than what we observed.

The main weakness of our cross‐sectional study is that a cause–effect relationship between the factors we examined and their effect on physical activity cannot be assumed. We have been careful to use the term “association” rather than “relationship”. Therefore, our calculation of the population effect of eliminating a risk factor, which assumes a cause–effect relationship, needs to be interpreted with caution. We make no claim here that making people feel safe in their neighbourhood would, in itself, increase the number of people who would be physically active. Rather, we have applied PIMs to highlight the potential effect of changes in particular neighbourhood factors on physical activity, and state that intervention studies are the only sure way to examine their effect. However, in practice, given the paucity of community‐based evaluations, policy makers often rely on cause–effect relationships to be assumed to some degree. We have merely applied a population perspective to such interpretation.

Few studies have previously examined the influence of feelings of safety on physical activity, particularly in the UK. A small cross‐sectional study in England23 found that women were more likely to walk at least 15 min a week if they felt safe during the day. In the US, perceptions of safety for walking were associated with actual walking,24 and crime was perceived as more of a problem in socially deprived areas that also had low levels of physical activity.25 Similarly, Americans who perceived their neighbourhood as less than extremely safe were more than twice as likely to have no leisuretime physical activity, and those who considered it to be not at all safe were nearly three times as likely to have no leisuretime physical activity.26 However, in a Danish study, although participating in sports activities was inversely related to perceptions about the amount of police attention their neighbourhood received,27 it was not found to influence walking and cycling activities.

What this paper adds

  • Few studies have considered the wider determinants of health on levels of physical activity in the population.
  • Feeling safe in the home and out and about in the neighbourhood may have as large an effect on population levels of physical activity as factors such as access to leisure facilities.

Policy implications

  • Making people feel safer in their neighbourhood is a key priority to increase population levels of physical activity.

Evidence on the possible role of perceptions relating to the location of leisure facilities and physical activity is conflicting. Our own findings support an independent association of perceived access to recreational facilities and physical activity, although its population effect was much less than for feelings of safety. This differs to the earlier study in England,23 but supports findings in Australia28 and the US.29,30 Consequently, we argue the urgent need to carry out prospective studies in the UK, which, wherever possible, will make full use of the many “natural experiments” around the country to obtain reliable evidence on the effect of contextual changes on population levels of physical activity.

Conclusion

Our study suggests that feeling unsafe in the neighbourhood is as much of a barrier to physical activity as how well people thought their home was for access to leisure facilities. As such, strategies to increase physical activity need to emphasise the perceived effect of feeling safe among the local population. Encouraging people to spend more time walking for leisure and commuting purposes seems to be a sensible approach to incorporate physical activity within activities of daily living. For this to become a reality, we need to start by ensuring that people feel safe out and about in their neighbourhood.

Abbreviations

PIM - population impact measure

Footnotes

Competing interests: None declared.

References

1. Department of Health Health survey for England 2004. Updating of trends data to include 2004 data London, DoH, 2005. http://www.ic.nhs.uk/pubs/hlthsvyeng2004upd (accessed 28 Apr 2006)
2. Harrison R A, McElduff P, Edwards R. Planning to win: health and lifestyles associated with physical activity amongst 15423 adults. Public Health 2006. 120206–212.212 [PubMed]
3. Vaz de Almeida M D, Grace P, Afonso C. et al Physical activity levels and body weight in a nationally representative sample in the European Union. Public Health Nutr 1999. 2105–113.113 [PubMed]
4. World Health Organization World health report. Geneva: WHO, 2002
5. Cavill N, Buxton K, Bull F. et alPromotion of activity among adults: evidence into practice briefing. London: National Institute for Health and Clinical Excellence, 2006
6. Lawlor D A, Ness A R, Cope A M. et al The challenges of evaluating environmental interventions to increase population levels of physical activity: the case of the UK National Cycle Network. J Epidemiol Community Health 2003. 5796–101.101 [PMC free article] [PubMed]
7. Berrigan D, Troiano R P. The association between urban form and physical activity in U.S. adults. Am J Prev Med 2002. 23(2 Suppl)74–79.79 [PubMed]
8. Hillsdon M, Foster C, Naidoo B. et al The effectiveness of public health interventions for increasing physical activity among adults: a review of reviews. In: Evidence briefing. London: Health Development Agency, 2004
9. Boslaugh S E, Luke D A, Brownson R C. et al Perceptions of neighbourhood environment for physical activity: is it “Who you are” or “Where you live”? J Urban Health 2004. 81671–681.681 [PMC free article] [PubMed]
10. Heller R. Evidence for population health. Oxford: Oxford Medical Publications, 2005
11. Heller R F, Buchan I, Edwards R. et al Communicating risks at the population level: application of population impact numbers. BMJ 2003. 3271162–1165.1165 [PMC free article] [PubMed]
12. Office for National Statistics 2005. http://neighbourhood.statistics.gov.uk (accessed 27 Jun 2005)
13. Harrison R A, Holt D, Elton P J. Do postage stamps increase response rates to postal health surveys? A randomised controlled trial. Int J Epidemiol 2002. 31872–874.874 [PubMed]
14. Office for National Statistics Census 2001. General report for England and Wales. London: Palgrave Macmillan, 2005
15. Townsend P, Phillimore P, Beattie A. Health and deprivation: inequality and the North. London: Croom Helm, 1988
16. Adams J, Ryan V, White M. How accurate are Townsend Deprivation Scores as predictors of self‐reported health? A comparison with individual level data. J Public Health 2004. 27101–106.106 [PubMed]
17. Godin G, Shephard R. A simple method to assess exercise behavior in the community. Can J Appl Sports Sci 1985. 10141–146.146 [PubMed]
18. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004. 159702–706.706 [PubMed]
19. Hanley J A. A heuristic approach to the formulas for population attributable fraction. J Epidemiol Community Health 2001. 55508–514.514 [PMC free article] [PubMed]
20. Gemmell I, Heller R F, McElduff P. et al Population impact of stricter adherence to recommendations for pharmacological and lifestyle interventions over one year in patients with coronary heart disease. J Epidemiol Community Health 2005. 591041–1046.1046 [PMC free article] [PubMed]
21. Hillsdon M, Thorogood M, Murphy M. et al Can a simple measure of vigorous physical activity predict future mortality? Results from the OXCHECK study. Public Health Nutr 2003. 7557–562.562 [PubMed]
22. Hill A, Roberts J, Ewings P. et al Non‐response bias in a lifestyle survey. J Public Health 1997. 19203–207.207 [PubMed]
23. Foster C, Hillsdon M, Thorogood M. Environmental perceptions and walking in English adults. J Epidemiol Community Health 2004. 58924–928.928 [PMC free article] [PubMed]
24. Li F, Fisher J, Brownson R C. et al Multilevel modelling of build environment characteristics related to neighbourhood walking activity in older adults. J Epidemiol Community Health 2005. 59558–564.564 [PMC free article] [PubMed]
25. Wilson D K, Kirtland K A, Ainswort B E. et al Socioeconomic status and perceptions of access and safety for physical activity. Ann Behav Med 2004. 2820–28.28 [PubMed]
26. Morbidity Mortality Weekly Reports Perceptions of neighbourhood characteristics and leisure‐time physical inactivity—Austin/Travis County, Texas, 2004. MMWR 2005. 54926–928.928 [PubMed]
27. Van Lenthe F J, Brug J, Mackenback J P. Neighbourhood inequalities in physical inactivity: the role of neighbourhood attractiveness, proximity to local facilities and safety in the Netherlands. Soc Sci Med 2005. 60763–775.775 [PubMed]
28. Ball K, Bauman A, Leslie E. et al Perceived environmental aesthetics and convenience and company are associated with walking for exercise among Australian adults. Prev Med 2001. 33434–440.440 [PubMed]
29. Booth M L, Owen N, Bauman A. et al Social‐cognitive and perceived environmental influences associated with physical activity in older Australians. Prev Med 2000. 3115–22.22 [PubMed]
30. Parks S E, Housemann R A, Brownson R C. Differential correlates of physical activity in urban and rural adults of various socio‐economic backgrounds in the United States. J Epidemiol Community Health 2003. 5729–35.35 [PMC free article] [PubMed]

Articles from Journal of Epidemiology and Community Health are provided here courtesy of BMJ Group
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • Cited in Books
    Cited in Books
    PubMed Central articles cited in books
  • MedGen
    MedGen
    Related information in MedGen
  • PubMed
    PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...