Temporal trends in physical activity in England: The Health Survey for
England 1991 to 2004
Emmanuel Stamatakisa,⁎, Ulf Ekelundb, Nicholas J. Warehamb
aDepartment of Epidemiology and Public Health, 1-19 Torrington Place, WC1E 6BT, London, UK
bMedical Research Council Epidemiology Unit, University of Cambridge, UK
Available online 8 January 2007
Objectives. Physical activity is an established risk factor for chronic disease but very little is known about its temporal trends in England. Such
information is crucial for planning public health interventions.
Methods. We explored temporal trends in occupational activity, walking, domestic activity, and sports using Health Survey for England data in
95,342 adults aged 16 and over. Data were collected annually in 1991–4, 1997–9, and 2003–04. Multivariate logistic regression and multiple
linear regression models assessed trends in physical activity for dichotomous and continuous outcomes, respectively. Analyses were adjusted for
age and social class.
Results. Physical activity levels at work declined over time but there was a consistent and significant upward trend in regular sports
participation among all age groups. Changes in questions in 1997 and 1999 confounded trends in walking and heavy domestic activity and total
physical activity. Between 1999 and 2004 (when physical activity questions remained unchanged), there were significant increases in average time
spent in all activity types and the percentage of adults meeting the current physical activity recommendations. These short-term increases were
more marked among adults aged 35 to 64.
Conclusion. The common perception that overall physical activity levels are declining may be over-simplistic as despite the decreases in
occupational physical activity, there is a clear upward trend in sports participation. Changes in the measuring methodology over time preclude the
presentation of a clear picture of the total temporal trends in physical activity in England.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Physical activity; Exercise; Trends; Obesity; England
Physical inactivity is associated with increased risk for all-
cause mortality (Lee and Skerrett, 2001) and many chronic
conditions, including ischemic heart disease (Kohl, 2001),
diabetes (Lynch et al., 1996), certain types of cancer (Thune and
Furberg, 2001), and obesity (United States Department of
Health and Human Services, 1996; Chakravarthy et al., 2002;
Chief Medical Officer, 2004). Inactivity in England is estimated
to cost the National Health Service over £8.2 billion a year
(Department of Culture, Media and Sports, 2002).
Despite the wide recognition that physical inactivity is a
major cause for chronic disease, no dedicated surveillance is in
place in the UK. The insufficiency of the existing evidence has
been highlighted in the first independent evidence-based
assessment of the long-term resource requirements for the
National Health Service (Wanless, 2004). Information on
temporal physical activity trends can identify population
subgroups at high risk for physical inactivity, evaluate public
health interventions, enhance our understanding on dose–
response issues, and help develop population specific physical
activity interventions (Macera and Pratt, 2000).
The rapidupwardobesityprevalence(Health andSocialCare
Information Centre, 2005) and the lack of a clear trend in dietary
intake (Prentice and Jebb, 1995; Stamatakis, 2006) may imply
that theoverallphysicalactivity energyexpenditure levels ofthe
population are declining. However, evidence supporting this
hypothesis is based on ecological proxy measures, such as
increases in car usage (Department of Transport, 2006) and TV
Available online at www.sciencedirect.com
Preventive Medicine 45 (2007) 416–423
E-mail address: firstname.lastname@example.org (E. Stamatakis).
0091-7435/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
measures or reports of physical activity. The only nationally
representative, population-based resource that has been collect-
ing multiple-domain physical activity data is the Health Survey
for England (HSfE), running continuously since 1991. No
published source has previously utilised all HSfE data to
investigate population physical activity temporal trends.
The aim of this study was, therefore, to examine the temporal
trends of participation in physical activity and to calculate the
proportion of adults meeting the current public health recom-
mendation for being active for at least 30 min per day on at least
5 days a week at a moderate intensity level (e.g. brisk walking,
intense domestic, or sports).
The HSfE is a continuous survey that since 1991 has drawn an annual
nationally representative sample of the general population living in households
(up to 10 adults per household). The sample is drawn using multi-stage stratified
probability sampling with postcode sectors as the primary sampling unit and the
Postcode Address File as the sampling frame for households. The adults'
physical activity module was included in the following surveys: 1991/2, 1993,
1994, 1997, 1998, 1999, 2003, and 2004.
Ethical approval had been granted by the London Medical Research Ethics
Council and/or Local Research Ethics Councils prior to each annual data
collection cycle of the Health Survey for England.
Data were collected by trained interviewers during household visits
throughout the year. Interviews enquired about personal socioeconomic and
demographic characteristics and physical activity participation in the four weeks
prior to the interview. Occupational (social) class was determined using the
Registrar-General's classification. Occupational physical activity was assessed
by a question on the perceived levels of physical activity at work (very/fairly
active, not very/not at all active). Frequency of participation in non-occupational
physical activity was examined as occasions (1991 to 1994) or days (1997 to
2004) respondents took part in heavy domestic (heavy housework, heavy DIY/
gardening), walking, and sports. For sports, respondents were shown a list of 10
rugby, callisthenics) and probed for participation in up to four additional
activities that were not listed. An additional question probed for the usual
duration of sports sessions. Intensity was assessed by asking respondents to
indicate if their usual walking pace was slow, average, fairly brisk, or fast and
whether sports and exercises made them feel out of breath or sweaty. Table 1
outlines the three different versions of the HSfE physical activity questionnaire
used between 1991 and 2004. The validity of the measuring instrument is
supported by the results of an unpublished study on 174 British adults where the
output of individually calibrated heart rate monitors (four times over a year, for
four consecutive days on each occasion) was compared against the early version
of the HSfE questionnaire. The HSfE questionnaire appeared to be a valid
measure of energy expenditure for occupational (p<0.05) and home (p<0.05)
activities in men and for the total physical activity score for men (p=0.05) and
women (p<0.05). A significant correlation was also observed between self-
reported activity from the questionnaire and aerobic fitness in men (p=0.001)
and women (p<0.05).
All heavy domestic physical activity was considered to be of moderate
intensity. Walking was classified as moderate physical activity if they reported
that their walking pace was “fairly brisk” or a “fast pace”, otherwise as light
intensity activity. All occasions of running/jogging, squash, cycling, aerobics,
keep fit, gymnastics, weight training, football/rugby, swimming, tennis/
badminton, and other sports of similar MET intensity (Ainsworth et al., 2000)
were coded as at least moderate intensity. Activities that could be either light or
moderate (Ainsworth et al., 2000), such as dancing, were coded as moderate
only if they made the subject out of breath or sweaty.
Data handling and analysis
We calculated the percentage of respondents aged 16–64 who were active at
work, defined as being in a manual occupation and reporting being fairly/very
active at work. We calculated moderate to vigorous physical activity (MVPA)
time for each physical activity type and percentage meeting the current
recommendations for participation in MVPA on at least 5 days per week for
30 min or more per day (Chief Medical Officer, 2004; United States Department
of Health and Human Services, 1996) including only non-occupational physical
activity types. Since the earliest (1991–94) and latest (1999–2004) versions of
the questionnaire had no duration questions on walking and domestic physical
activity sessions, we extracted the sex and age (5-year bands) mean session
duration values from the 1997–98 datasets and assigned the resulting values to
each respondent to calculate time in MVPA.
Trends in sex, age, and social class across HSfE year were assessed by Chi-
square tests. Trends in occupational physical activity and percentage classified
as meeting the current physical activity recommendations over time were
assessed using logistic regression where year was the independent variable.
Trends in sports time and total MVPA time across HSfE years were examined
using linear regression, adjusted for social class and age when necessary. Linear
regression analyses for heavy domestic, brisk to fast walking, and total MVPA
time were repeated for the years between 1999 and 2004 to determine short-term
trends over a period when physical activity questions remained unchanged.
Outline of the three versions of the Health Survey for England physical activity
Physical activity domain 1991–1994 1997–19981999–2004
Heavy domestic (heavy housework and heavy gardening/manual/building/DIY)
Lower limit for duration (minutes) None
Lower limit for duration (minutes)
Sports and exercise
Lower limit for duration (minutes)
Lower limit for duration (minutes)
a1991-94: occasions of participation; 1997-04: days of participation.
bA question probing on second walking session on the same day was
cA single question on overall activity levels at work.
417E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
Demographics and response rates
In total 95,342 HSfE respondents (43,401 men) completed
the physical activity interviews between 1991 and 2004 and
were included in the analysis. As Table 2 shows, response rates
04). There were significant increases in mean age (p<0.001),
ratio of non-manual over manual occupations (p<0.001), and
ratio of women over men (p<0.001) between 1991/2 and 2004.
Occupational physical activity
were active at work between 1991/2 and 2004. Significant
declines in occupational activity between 1991/2 and 2004 were
recorded for both men (from 43.4% to 38.5%, age-adjusted p<
Looking at narrow age groups, changes reached statistical sig-
nificance in men aged 16–34 (p<0.05), men 35–49 (p<0.05),
and women aged 50 to 64 (p<0.001). In most age groups the
decline was most pronounced from 1998 and onwards.
As Fig. 2 shows, mean time spent in sports of moderate to
vigorous intensity increased steadily among all age groups
(occupational class and age-adjusted p<0.001). An analogous
pattern was found when regular participation (at least once a
week) in sports and exercise was considered (p<0.001). The
adjusted odds of regular participation in moderate to vigorous
sports in 2004 was 1.74 for men (95%CI: 1.54 to 1.96) and 1.80
in women (95%CI: 1.61 to 2.01), compared with 1991/2. The
upward trend in regular sports participation was evident in all
age groups but particularly pronounced among the middle aged
and older groups (i.e. aged 35 and over). For example, the
percentage of those aged 35 to 49 participating regularly in
sports rose from 26.0% in 1991/2 to 42.6% in 2004 (p<0.001)
among men and from 25.1 to 35.4 among women (p<0.001).
Walking and heavy domestic activity
Trends in walking and domestic activity were less clear than
trends in occupational activity and sports due to the changes in
for both physical activity types change abruptly in those years
when the relevant questions were modified (1997 and 1999).
Between 1999 and 2004 (when questions remained unchanged)
the average weekly time spent in moderate-intensity walking
increased from 50 to 60 min for men (p<0.001) and from 38 to
50 min in women (p<0.001). Similarly, average weekly time
Total time spent at MVPA (p<0.001 for both men and
women) and the percentage of adults meeting the current
physical activity recommendations increased significantly for
Sample demographic characteristics and response rates by survey year: the Health Survey for England 1991 to 2004
1991/2 19931994 1997 1998 199920032004pTrend
0.86 0.91 0.830.830.83 0.840.80 0.75<0.001
Mean age±SD (years)
Age group (N)
Social class (%)
Response rates (%)
Overall response 797672 71697066 66<0.001
418E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
men (p<0.005) and women (p<0.001) over the 1999–2004
period. For all adults there was an increase of 1.7 percentage
points in the percentage meeting the recommended levels of
physical activity (from 46.8% to 48.5%) with little change in the
16–34 and over 65 age groups but clear increases among those
aged 35–49 (from 49.1% in 1999 to 55.2% in 2004, p<0.001)
and 50–64 years (from 42.9% to 47.1%, p<0.001).
This is, to our knowledge, the first study to examine
temporal trends across multiple physical activity domains in
English men and women over approximately the last 15 years,
a period when the prevalence of obesity in adults living in
England rose by over 65% (Health and Social Care Information
Centre, 2005). We found a consistent upward trend for regular
participation and time spent in sports but a decline in
occupational physical activity between 1991/2 and 2004. The
results also suggested a consistent upward short-term trend in
overall physical activity participation and in the percentage of
adults meeting the physical activity recommendations, between
1999 and 2004.
The key strengths of this study are the large, nationally
representative samples, the relatively high response rates, and
Fig. 2. Mean weekly time spent in moderate to vigorous-intensity sport activities lasting for at least 15 min. Health Survey for England 1991/2 to 2004. *Occupational
class-adjusted p for trend <0.001.
Fig. 1. Percentage of men and women aged 16–64 with high activity levels at work.†Health Survey for England 1991/2 to 2004.†Defined as being in a manual
occupation and reporting being fairly/very active at work.*Age and occupational class-adjusted p for trend <0.05. **Age and occupational class-adjusted p for trend
<0.001. ***Age-adjusted p for all men's 1999 to 2004 trend <0.01.
419 E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
the availability of data on all key physical activity domains. The
convergent (“face”) validity of the physical activity question-
naire has been indicated by its ability to clearly grade
anthropometric and biological outcomes, such as body mass
index, resting pulse rate, and HDL cholesterol, according to
physical activity levels (Stamatakis et al., in press). However,
no published data on its criterion validity are available.
Limitations include the reliance on self-report and the changes
in questions over time that make trends data in two key physical
activity domains (walking and domestic activity) and overall
activity levels difficult to interpret. Finally, the lack of
information on non-cooperating households may bias the
The percentage of those reporting high physical activity
levels at work decreased, a decrease more pronounced from
1998 and onwards. The overall decrease in manual occupations
(Table 2) may reflect advances in automation and computerisa-
tion of employment in the recent years. Since our definition of
high occupational physical activity involves occupational class,
differential response rates between social classes could be a
potential source of bias. People from lower socioeconomic
strata are more likely to be non-responders in survey research
(Department of Health, 1995; Turrell et al., 2002) but there is no
indication that response rates decreased more in manual than in
non-manual classes over time. As a consequence, our results
may underestimate the prevalence of high occupational physical
activity at any survey year point but it is unlikely that the
observed temporal trend data are affected.
Increases in time spent in sports were evident in all age
groups but particularly pronounced among the middle and late
life age groups (aged 35 to 49 and older). Similar findings
have also been reported among a representative sample of
middle-aged and older Scottish adults in a survey using a
similar to the HSfE methodology (Scottish Executive, 2005).
Fig. 4. Mean weekly time spent in heavy domestic physical activity, by year. Health Survey for England 1991/2 to 2004. *Age and occupational class-adjusted p for
trend <0.001. **Age and occupational class-adjusted p for 1999 to 2004 trend <0.001.
Fig. 3. Mean weekly time spent in brisk to fast walking, by survey year. Health Survey for England 1991/2 to 2004. *Age and occupational class-adjusted p for 1991/2
to 2004 trend <0.001. **Age and occupational class-adjusted p for 1999 to 2004 trend <0.001.
420 E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
When we examined temporal trends for specific activities,
results indicated a decline in cycling over time, whereas other
activities, such as weight training, fitness or health club
exercises, and general strength exercises such as press-ups and
push-ups increased considerably. Market research data support
this observation as the percentage of UK adults visiting
fitness, sports, and leisure facilities on a weekly basis
increased from 10% in 1991 to 14% in 2004, while regular
visitors (more than once a week) increased from 6% to 13%
over the same period (Mintel International Group Ltd., 2004).
These findings suggest that the commonly held view that
physical activity levels of the population are declining
(Prentice and Jebb, 1995) may be over-simplistic. This view,
often cited in public health (Chief Medical Officer, 2004),
health statistics (Stamatakis, 2006), and health policy (Health
Select Committee, 2004) reports, is based on proxy ecological
evidence (Department of Transport, 2006; Rickards et al.,
2004). This type of data does not take into account that
temporal changes in a certain domain (e.g. reduction in miles
walked) may be counter-balanced by changes in the opposite
direction in another domain, for example, an increase in sports
Unfortunately, our data are not sufficient to provide a
definitive answer as to whether overall physical activity has
changed over the last 15 years. As Fig. 5 indicates, fluctuations
in time spent at MVPA and the percentage of men and women
meeting the physical activity recommendations are more likely
a function of changes in the questions included in different
surveys and the imposition of certain analytical assumptions
these changes forced, rather than a reflection of true changes in
the physical activity levels of the population. Changes in the
wording of physical activity questions may have a considerable
impact on prevalence estimates (Welk, 2002).
For example, while there was no lower time limit for
heavy domestic activities in 1991 to 1994, a 15-minute limit
was introduced in 1997 and frequency of all physical activity
types was assessed as occasions in 1991–94 but as number of
days from 1997 onwards. Our attempt to extrapolate walking
and domestic activity duration using 1997/8 data may have
introduced additional error in our analyses. These measure-
ment and analytical limitations are reflected in Fig. 4 as a
sharp drop in self-reported heavy domestic activity time in
1997. These activities are highly prevalent (Lawlor et al.,
2002; Phongsavan et al., 2004; Stamatakis et al., in press) and
this change may have added considerable bias on the
frequency-based estimates from 1997 onwards. The purpose
for these changes in the questionnaire in 1997 is unclear as
the current physical activity recommendations for public
health were already in place (United States Department of
Health and Human Services, 1996; Department of Health,
1999) and the upward obesity trend had already been
documented (Food Standards Agency, 1990; Department of
Although the time period between 1999 and 2004 may be
too short for a temporal trend to emerge, we found significant
increases in all non-occupational physical activity domains
and the percentage of adults meeting the physical activity
recommendations. Increases in recreational exercise participa-
tion between 1989 and 2000 and increases in the prevalence of
walking between 1987 and 2000 have also been found among
Australian (Merom et al., 2006) and American adults
(Simpson et al., 2003), respectively. Such trends are
paradoxical, given that adult obesity prevalence has increased
worldwide (James et al., 2001), percentage points and
available data suggest that the population dietary intake has
remained stable (Stamatakis, 2006). The explanation of such
phenomena may lie in reporting bias associated with the
population health-related awareness, the publicity the physical
activity recommendations have received, or other more general
social trends. The rapid upward obesity trend (Health and
Fig. 5. Total weekly time spent in moderate to vigorous physical activity and percentage of adults meeting the physical activity recommendations. The Health Survey
for England 1991/2 to 2004. *Age and occupational class-adjusted p for 1999 to 2004 trend <0.001. **Age and occupational class-adjusted p for 1999 to 2004 trend
421E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
Social Care Information Centre, 2005) has triggered a
dramatic increase in media coverage (International Food
Information Council Foundation, 2006) and “reality televi-
sion” programmes on issues concerning fitness, exercise, diet,
and weight loss. These phenomena and the publicity the
physical activity recommendations (United States Department
of Health and Human Services, 1996; Chief Medical Officer,
2004; Merom et al., 2006) have received may have increased
the effect of social desirability bias on our estimates in more
recent years, contributing to the upward trend observed.
Regardless of the resources spent and efforts made to develop
meticulously any self-reported physical activity instruments,
such sources of systematic measurement error cannot be
A possible solution to the problems related to self-reported
physical activity measures would be the introduction of
objective physical activity measures, such as accelerometers,
in population surveys. Objective physical activity monitoring
could supplement self-reported data (e.g. physical activity
mode) with far more accurate information on the frequency,
duration, and intensity of free-living physical activity. The
2003/04 National Health and Nutrition Examination Survey
(NHANES) in the US has shown that objective physical
activity monitoring is feasible in large scale nationwide
population studies. Introducing objective monitoring of
physical activity should be part of any physical activity
surveillance system, such as the Health Survey for England.
Our results suggest that participation in sports has increased
whereas occupational physical activity may have decreased in
English adults over the last 15 years. However, these results
provide a partial picture of physical activity temporal trends in
England between the early 1990s and 2004 as highly prevalent
activities such as walking have been inconsistently measured
over time. There is an urgent need for a carefully designed UK
physical activity surveillance system that includes objective
measures of physical activity.
The Health Survey for England was funded by the English
Department of Health. The data presented here are available
from the UK Data Archive (http://www.data-archive.ac.uk/
Ainsworth, B.E., Haskell, W.L., Whitt, M.C., Irwin, M.L., Swartz, A.M.,
Strath, S.J., et al., 2000. Compendium of physical activities: an update of
activity codes and MET intensities. Med. Sci. Sports Exercise 3,
Chakravarthy, M.V., Joyner, M.J., Booth, F.W., 2002. An obligation for
primary care physicians to prescribe physical activity to sedentary patients
to reduce the risk of chronic health conditions. Mayo Clin. Proc. 77,
Chief Medical Officer, 2004. At Least Five AWeek: Evidence on the Impact of
Physical Activity and Its Relationship to Health. Department of Health,
Department of Culture, 2002. Media and Sports. Game Plan: A Strategy for
Delivering the Government's Sport and Physical Activity Objectives.
Strategy Unit, London, England.
Department of Health, 1993. The Health Survey for England 1991. The
Stationery Office, London, England.
Department of Health, 1995. The Health Survey for England 1994. Volume II:
Survey Methodology and Documentation. The Stationery Office, London,
Department of Health, 1999. The Health Survey for England 1998: Cardio-
vascular Disease. The Stationery Office, London, England.
Department of Transport, 2006. Transport Trends: 2005 Edition. Department of
Transport, London, England. Available online: www.dft.gov.uk/stellent/
(accessed 06 July 2006).
Food Standards Agency, 1990. The Dietary and Nutritional Survey of British
Adults. The Stationery Office, London, England.
Health and Social Care Information Centre, 2005. Health Survey for England
2004. Updating of Trend Tables to Include 2004 Data. Health and Social
Care Information Centre, London, England. Available online: http://www.ic.
nhs.uk/pubs/hlthsvyeng2004upd (accessed 06 July 2006).
Health Select Committee, 2004. Third Report of Session 2003–04, Obesity
HC23–1. The Stationery Office, London, England.
International Food Information Council Foundation. Trends in Obesity-
related media coverage. Available online: http://www.ific.org/research/
obesitytrends.cfm (accessed 06 July 2006).
James, P.T., Leach, R., Kalamara, E., Shayeghi, M., 2001. The worldwide
obesity epidemic. Obes. Res. 9, S228–S233.
Kohl III, H.W., 2001. Physical activity and cardiovascular disease: evidence for
a dose response. Med. Sci. Sports Exercise 33, S472–S483.
Lawlor, D.A., Taylor, M., Bedford, C., Ebrahim, S., 2002. Is housework good
for health? Levels of physical activity and factors associated with activity in
elderly women. Results from the British Women's Heart and Health Study.
J. Epidemiol. Commun. Health 56, 473–478.
Lee, I.-M., Skerrett, P.J., 2001. Physical activity and all-cause mortality: what is
the dose-response relation? Med. Sci. Sports Exerc. 33, S459–S471.
Lynch, J., Helmrich, S.P., Lakka, T.A., Kaplan, G.A., Cohen, R.D., Salonen, R.,
Salonen, J.T., 1996. Moderately intense physical activities and high levels of
cardiorespiratory fitness reduce the risk of non-insulin-dependent diabetes
mellitus in middle-aged men. Arch. Intern. Med. 156, 1258–1354.
Macera, C.A., Pratt, M., 2000. Public Health Surveillance of physical activity.
Res. Q. Exerc. Sport 71, 97–103.
Merom, D., Phongsavan, P., Chey, T., Bauman, A., 2006. Long-term changes
in leisure time walking, moderate and vigorous exercise: were they in-
fluenced by the national physical activity guidelines? J. Sci. Med. Sport 9,
Mintel International Group Ltd., 2004. Leisure Centres and Swimming Pools—
UK. Mintel International Group Ltd, London, England.
National Health and Nutrition Examination Survey, 2003–04. Laboratory
Procedures Manual, Chapter 16: Physical Activity Monitor. Available
(accessed 06 July 2006).
Phongsavan, P., Merom, D., Marshall, A.L., Bauman, A., 2004. Estimating
physical activity levels: the role of domestic activities. J. Epidemiol.
Commun. Health 58, 466–467.
Prentice, A.M., Jebb, S.A., 1995. Obesity in Britain: gluttony or sloth? Br. Med.
J. 311, 437–439.
Rickards, L., Fox, K., Roberts, C., Fletcher, L., Goddard, D., 2004. Living in
Britain: Results from the 2002 General Household Survey. Office for
National Statistics, London, England.
Scottish Executive, 2005. The Scottish Health Survey 2003. The Scottish
Simpson, M.E., Serdula, M., Galuska, D.A., Gillespie, G., Donehoo, R.,
Macera, C., Mack, K., 2003. Walking trends among U.S. adults. The
Behavioral Risk Factor Surveillance System, 1987–2000. Am. J. Prev. Med.
422E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423
Stamatakis, E., 2006. Obesity, eating, and physical activity. In: Bajekal, M., Download full-text
Osborne, V., Yar, M., Meltzer, H. (Eds.), Focus on Health. Office for
National Statistics/PalGrave Macmillan, London, England, pp. 48–61.
Stamatakis, E., Hillsdon, M., Primatesta, P., in press. Domestic physical activity
in relation to multiple cardiovascular disease risk factors. Am. J. Prev. Med.
Thune,I., Furberg, A.-S.,2001.Physicalactivity andcancerrisk: dose–response
and cancer, all sites and site-specific. Med. Sci. Sports Exercise 33,
Turrell, G., Patterson, C., Oldenburg, B., Gould, T., Roy, M.-A., 2002. The
socio-economic patterning of survey participation and non-response error in
a multi-level study of food purchasing behaviour: area- and individual-level
characteristics. Public Heath Nutr. 6, 181–189.
United States Department of Health and Human Services, 1996. Physical
Activity and Health: A Report of the Surgeon General. Centres for Disease
Control, Atlanta, GA.
Wanless, D., 2004. Securing Good Health to the Whole Population. Final report.
Department of Health, London, England.
Welk, G.J., 2002. Introduction to physical activity research. In: Welk, G.J. (Ed.),
Physical Activity Assessments for Health-related Research. Human
Kinetics, Champaign, IL.
423 E. Stamatakis et al. / Preventive Medicine 45 (2007) 416–423