Articles

Worldwide trends in insufficient physical activity from

2001 to 2016: a pooled analysis of 358 population-based
surveys with 1·9 million participants
Regina Guthold, Gretchen A Stevens, Leanne M Riley, Fiona C Bull

Summary
Background Insufficient physical activity is a leading risk factor for non-communicable diseases, and has a negative Lancet Glob Health 2018
effect on mental health and quality of life. We describe levels of insufficient physical activity across countries, and Published Online
estimate global and regional trends. September 4, 2018
http://dx.doi.org/10.1016/
S2214-109X(18)30357-7
Methods We pooled data from population-based surveys reporting the prevalence of insufficient physical activity,
See Online/Comment
which included physical activity at work, at home, for transport, and during leisure time (ie, not doing at least 150 min http://dx.doi.org/10.1016/
of moderate-intensity, or 75 min of vigorous-intensity physical activity per week, or any equivalent combination of the S2214-109X(18)30381-4
two). We used regression models to adjust survey data to a standard definition and age groups. We estimated time Department for Prevention of
trends using multilevel mixed-effects modelling. Noncommunicable Diseases,
WHO, Geneva, Switzerland
(R Guthold PhD, L M Riley MSc,
Findings We included data from 358 surveys across 168 countries, including 1·9 million participants. Global age- Prof F C Bull PhD); Department
standardised prevalence of insufficient physical activity was 27·5% (95% uncertainty interval 25·0–32·2) in 2016, for Information, Evidence and
with a difference between sexes of more than 8 percentage points (23·4%, 21·1–30·7, in men vs 31·7%, 28·6–39·0, Research, WHO, Geneva,
Switzerland (G A Stevens DSc);
in women). Between 2001, and 2016, levels of insufficient activity were stable (28·5%, 23·9–33·9, in 2001; change not
and Department of Sport and
significant). The highest levels in 2016, were in women in Latin America and the Caribbean (43·7%, 42·9–46·5), Exercise Science, University of
south Asia (43·0%, 29·6–74·9), and high-income Western countries (42·3%, 39·1–45·4), whereas the lowest levels Western Australia, Perth, WA,
were in men from Oceania (12·3%, 11·2–17·7), east and southeast Asia (17·6%, 15·7–23·9), and sub-Saharan Africa Australia (Prof F C Bull)
(17·9%, 15·1–20·5). Prevalence in 2016 was more than twice as high in high-income countries (36·8%, 35·0–38·0) as Correspondence to:
in low-income countries (16·2%, 14·2–17·9), and insufficient activity has increased in high-income countries over Dr Regina Guthold, World Health
Organization, 1211 Geneva,
time (31·6%, 27·1–37·2, in 2001). Switzerland
gutholdr@who.int
Interpretation If current trends continue, the 2025 global physical activity target (a 10% relative reduction in
insufficient physical activity) will not be met. Policies to increase population levels of physical activity need to be
prioritised and scaled up urgently.

Funding None.

Copyright This is an Open Access article published under the CC BY 3.0 IGO license which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.  In any use of this article,
there should be no suggestion that WHO endorses any specific organisation, products or services. The use of the
WHO logo is not permitted. This notice should be preserved along with the article’s original URL.

Introduction to assess the effectiveness of policy, and guide future

The health benefits of physical activity are well policy and programme planning.
established and include a lower risk of cardiovascular The first compilation of country data to produce global
disease, hypertension, diabetes, and breast and colon and regional estimates of insufficient physical activity
cancer. Additionally, physical activity has positive effects was undertaken in the early 2000s, as part of the Global
on mental health, delays the onset of dementia, and can Burden of Disease study.7 Bull and colleagues7 included
help the maintenance of a healthy weight.1–5 data for physical activity from 34 mainly high-income
In recognition of this strong link between physical countries, mostly focusing on leisure time physical
activity and major non-communicable diseases, member activity. Activity performed in other domains (activity at
states of WHO agreed to a 10% relative reduction in the work, in the household, and for transport) had to
prevalence of insufficient physical activity by 2025, as one be estimated for most countries, in order to get
of the nine global targets to improve the prevention and comprehensive and comparable results. Subsequently,
treatment of non-communicable diseases.6 Monitoring two questionnaires including all activity domains were
current levels and trends of insufficient physical activity developed: the International Physical Activity Question­
is essential to track progress towards this global physical naire8 and the Global Physical Activity Questionnaire.9,10
activity target, but also to identify high-risk populations, Both questionnaires have since been used in many

www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7 1

 Articles

Research in context
Evidence before this study Global age-standardised prevalence of insufficient physical
WHO produced the first set of internationally comparable activity was 27·5% (95% uncertainty interval 25·0–32·2) in
estimates on insufficient physical activity in 2008 for 2016. The prevalence has been stable since 2001.
122 countries. These estimates were updated for 2010, and
Implications of all the available evidence
published in the 2016 Lancet Series on Physical activity for
Progress towards achieving the global target of a 10% relative
146 countries, showing a global prevalence of insufficient
reduction of insufficient physical activity by 2025, has been too
physical activity of 23·3%, with higher levels among women and
slow. Accelerated action is needed to reverse trends in central
older age groups. However, no compilation of global data on
and eastern Europe, high-income Western countries, Latin
adult physical inactivity has been undertaken since, and no
American and the Caribbean, and south Asia. Policies and
regional and global trends have previously been developed
programmes are also needed to achieve or maintain low levels
because data for trends have been considered too scarce.
of inactivity in other regions and in lower-income countries.
Added value of this study Implementation of targeted evidence-based interventions
This study provides the most complete description of global, presented in the Global Physical Activity Action Plan 2018–2030
regional, and country levels of insufficient physical activity and, will improve population health and help deliver many of the
for the first time, presents regional and global trends over time. 2030 Sustainable Development Goals.

surveys, including several large, inter­ national surveys walking and cycling), and during leisure time (ie, sports
such as the World Health Survey,11 the Eurobarometer and active recreation); (2) data were collected through
surveys,12 and the WHO Stepwise Approach to NCD Risk random sampling with a sample size of at least 200, and
Factor Surveillance.13,14 Mainly based on data from these were representative of a national or defined subnational
questionnaires, WHO produced comparable estimates of population; (3) prevalence of insufficient physical activity
insufficient physical activity in 122 countries in 2008, and was reported by age and sex, according to current WHO,3
updated them for 2010 for 146 countries. Both sets of or former physical activity recommendations.21 These
estimates were published in The Lancet Physical Activity former guidelines recom­mended for adults to spread out
Series4,15 and used to calculate the effect of insufficient their activity during the week in doing at least 30 min of
physical activity on non-communicable diseases.16 moderate activity on at least 5 days, or equivalent. They
However, in The Lancet Physical Activity 2016 Series, were in place until an update of evidence in 2008,1 when
Sallis and colleagues4 noted that consistent data for this requirement was removed.
trends in adult physical activity are still scarce. As a Physical activity data collected using wearable devices,
consequence, no article has reported regional and global such as accelerometers or pedometers, were not included
trends of insufficient physical activity. We updated because of the limited comparability with self-reported
previously published4,15,17 country, regional, and global data. Where available, we used individual-level data to
estimates of adult prevalence of insufficient physical calculate the prevalence of insufficient physical activity,
activity with new data and new methods, and estimated, taking the sampling designs into account. Where raw
for the first time, global and regional trends from 2001 to data were not available, we used aggregated data as
2016. reported. We included all data that met the inclusion
criteria and that were provided before the end of
Methods September, 2017.
Study design We obtained data from WHO and other international
We estimated the prevalence of insufficient physical surveys.11–14 We also did a systematic literature search of
activity in adults aged 18 years and older, in 168 countries, PubMed, up to October, 2015, for articles published in
for three World Bank income groups,18 nine regions English, Portuguese, German, Spanish, Chinese, and
See Online for appendix (appendix p 1),19,20 and globally for 2001–16. Insufficient Italian, including the International Physical Activity
physical activity was defined as adults not meeting Question­ naire and the Global Physical Activity
the WHO recommendations3 on physical activity for Questionnaire as search terms. To identify additional
health—ie, at least 150 min of moderate-intensity, or data sources, we also analysed and verified results of the
75 min of vigorous-intensity physical activity per week, 2017 WHO NCD Country Capacity Survey,22 in which
or any equivalent combination of the two. each WHO member state answered a question on the
inclusion of physical activity in national risk factor
Data sources surveys. We also had personal communications with
We included data that fulfilled the following criteria: (1) the WHO regional focal points, personal networks, and
survey questionnaire explicitly included physical activity directly with researchers, including inquiries about
across four key domains—ie, for work, in the household additional data from authors of published studies.
(paid or unpaid), for transport to get to and from places (ie, Finally, we had a 6-week consultation with all WHO

2 www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7

 Articles

member states in which countries commented on the Second, we adjusted for over-reporting in the
first draft of estimates and submitted any additional International Physical Activity Questionnaire. This
data. After the consultation, the estimates were updated question­ naire over-reports physical activity, leading to
during October, 2017, with new data that were submitted. an underestimation of the prevalence of insufficient
In total, 358 surveys from 168 countries were included in activity.23–26 To correct for this over-reporting, we applied
the analysis (appendix pp 2–15). an adjustment factor to surveys that used the Inter­
national Physical Activity Questionnaire. This adjustment
Statistical analysis factor was determined by exploring differences in
Survey data were sometimes not comparable because prevalence between the International Physical Activity
of differences in study design. We applied four key Questionnaire and other, similar survey questionnaires,
adjustments to survey data using linear regression such as the Global Physical Activity Questionnaire, in
modelling to improve comparability (appendix pp 17–27). countries that have used both.
First, we converted definitions. For surveys in which data Third, we adjusted data from surveys that had only
were reported only for the former recommendation on urban samples. Of the 358 surveys used, 27 reported data
physical activity for health,21 and not for the current for urban populations only. For these surveys, we
recommendation,3 we converted data to the current estimated the prevalence in rural areas using information
recommendation, our target indicator. from surveys reporting both urban and rural prevalence.

All Region Income group

countries
(n=231)
Central Asia, Central East and High-income High-income Latin Oceania South Sub-Saharan Low-income Middle- High-income
Middle East, and southeast Asia Pacific Western American (n=22) Asia Africa (n=31) income (n=75)
north Africa eastern Asia (n=3) countries and (n=6) (n=53) (n=109)
(n=28) Europe (n=16) (n=36) Caribbean
(n=20) (n=47)
Countries with 168 23 17 13 3 24 25 17 5 41 24 89 54
data (%) (72·7%) (82·1%) (85·0%) (81·3%) (100·0%) (66·7%) (53·2%) (77·3%) (83·3%) (77·4%) (77·4%) (81·7%) (72·0%)
Percentage of 95·8% 90·4% 98·3% 97·3% 100·0% 99·0% 87·6% 94·3% 98·5% 89·6% 81·9% 96·9% 96·3%
population
covered with data
Number with 65 8 11 7 2 20 5 7 1 4 1 27 36
trend data* (%) (28·1%) (28·6%) (55·0%) (43·8%) (66·7%) (55·6%) (10·6%) (31·8%) (16·7%) (7·5%) (3·2%) (24·8%) (48·0%)
*Defined as having used the same instrument and coverage at more than one point in time.

Table 1: Distribution of data across income groups and regions

Overall percentage of Percentage of men Percentage of Country with the minimum prevalence Country with the maximum prevalence
insufficient physical with insufficient women with
activity (95% UI) physical activity insufficient physical
(95% UI) activity (95% UI)
Country Prevalence (95% CI) Country Prevalence (95% CI)
All countries 27·5% (25·0–32·2) 23·4% (21·1–30·7) 31·7% (28·6–39·0) Uganda 5·5% (4·0–7·6) Kuwait 67·0% (58·6–74·3)
Central Asia, Middle East, and 32·8% (31·0–35·2) 25·9% (23·7–28·7) 39·9% (37·9–42·7) Jordan 11·9% (8·4–16·4) Kuwait 67·0% (58·6–74·3)
north Africa
Central and eastern Europe 23·4% (20·9–28·0) 22·0% (18·6–28·8) 24·7% (21·7–33·9) Moldova 11·5% (8·1–16·0) Serbia 39·5% (30·8–48·8)
East and southeast Asia 17·3% (15·8–22·1) 17·6% (15·7–23·9) 16·9% (14·9–25·7) Cambodia 10·5% (6·9–15·7) Philippines 39·7% (31·3–48·6)
High-income Asia Pacific 35·7% (34·4–37·0) 33·0% (29·4–33·6) 38·3% (37·4–42·6) South Korea 35·4% (20·9–52·9) Singapore 36·5% (21·7–54·3)
High-income Western 36·8% (34·6–38·4) 31·2% (28·5–32·6) 42·3% (39·1–45·4) Finland 16·6% (12·9–21·0) Cyprus 44·4% (36·8–52·1)
countries
Latin America and Caribbean 39·1% (37·8–40·6) 34·3% (32·5–35·5) 43·7% (42·9–46·5) Dominica 21·6% (16·3–28·0) Brazil 47·0% (38·9–55·3)
Oceania 16·3% (14·3–20·7) 12·3% (11·2–17·7) 20·3% (18·8–28·7) Niue 6·9% (4·8–9·9) American Samoa 53·4% (41·4–65·0)
South Asia 33·0% (23·0–51·7) 23·5% (14·4–54·3) 43·0% (29·6–74·9) Nepal 13·4% (11·2–15·6) India 34·0% (22·3–47·7)
Sub-Saharan Africa 21·4% (19·1–23·3) 17·9% (15·1–20·5) 24·8% (21·8–27·2) Uganda 5·5% (4·0–7·6) Mauritania 41·3% (33·4–49·2)
Low-income 16·2% (14·2–17·9) 13·4% (11·3–15·6) 18·8% (15·9–21·4) Uganda 5·5% (4·0–7·6) Mali 40·4% (33·6–47·3)
Middle-income 26·0% (22·6–31·8) 21·9% (18·9–31·3) 30·1% (26·0–39·5) Lesotho 6·3% (4·5–8·6) American Samoa 53·4% (41·4–65·0)
High-income 36·8% (35·0–38·0) 32·0% (29·8–33·1) 41·6% (39·1–43·9) Finland 16·6% (12·9–21·0) Kuwait 67·0% (58·6–74·3)

Table 2: Prevalence of insufficient physical activity in 2016

www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7 3

 Articles

Central Asia, Middle East, north Africa Central and eastern Europe East and southeast Asia
100

80
Prevalence (%)

60

40

20

High-income Asia Pacific High-income western countries Latin America and Caribbean
100

80
Prevalence (%)

60

40

20

Oceania South Asia Sub-Saharan Africa

100

80
Prevalence (%)

60

40

20

0
2000 2005 2010 2015 2000 2005 2010 2015
World Year Year
100

80
Prevalence (%)

60

40

20

0
2000 2005 2010 2015
Year

Figure 1: Trends in insufficient physical activity in men from 2001 to 2016

The shaded areas show 95% uncertainty intervals.

We calculated a national estimate combining the two by allow estimates to be informed by data from the same
applying estimates of population by area of residence for country, from other countries in the region, and other
the respective survey year.27 variables, this model included a random slope on year, a
Fourth, for surveys that did not report data for all ages random intercept for each country, and fixed effects for
older than 18 years, we estimated the data for the missing country urbanisation, education, and location within
age groups using information from surveys that reported nine previously defined regions that have been used in
age-specific values for the entire age range. similar analysis for other non-communicable disease
The methods used to generate final estimates of risk factors.19,20 For 76 countries, only one survey was
insufficient physical activity prevalence by country, year, available. We assumed no change over time in these
sex, and age differed depending on the availability of data countries, based on the fact that the average change in
on trends over time. prevalence per year across the 65 countries with at least
Of the 168 countries included, 65 had done at least two comparable surveys was less than 0·01%.
two comparable surveys from different years using the 27 countries had done several surveys with different
same questionnaire (appendix p 16). For these countries, survey coverage, or with different questionnaires,
we estimated the prevalence of insufficient physical limiting comparability of these survey data. Although we
activity for each year from 2001, to 2016, using a adjusted for over-reporting in the International Physical
multilevel mixed-effects linear regression model. To Activity Questionnaire, we considered the comparison of

4 www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7

 Articles

Central Asia, Middle East, north Africa Central and eastern Europe East and southeast Asia
100

80
Prevalence (%)

60

40

20

High-income Asia Pacific High-income western countries Latin America and Caribbean
100

80
Prevalence (%)

60

40

20

Oceania South Asia Sub-Saharan Africa

100

80
Prevalence (%)

60

40

20

0
2000 2005 2010 2015 2000 2005 2010 2015
World Year Year
100

80
Prevalence (%)

60

40

20

0
2000 2005 2010 2015
Year

Figure 2: Trends in insufficient physical activity in women from 2001 to 2016

The shaded areas show 95% uncertainty intervals.

adjusted data from the International Physical Activity 2·5th and 97·5th percentile of the 1000 draws. We
Questionnaire to data from other questionnaires to be calculated trends as the difference in prevalence
insufficient to inform country trends. Therefore, for between 2001, and 2016, and considered a change in
these 27 countries, we also assumed a flat trend, whereby prevalence over time to be statistically significant if
the trend line was based on the average prevalences fewer than 2·5% of draws showed changing trend
across surveys for each country. (appendix pp 17–27).30
To derive final estimates for the entire age range, we
age-standardised resulting sex-specific and age-specific Role of the funding source
prevalence estimates using the WHO Standard Popu­ There was no funding source for this study. The
lation.28 To produce global and regional estimates, corresponding author had full access to all the data and
and estimates for World Bank income groups,18 we had final responsibility for the decision to submit
created population-weighted sex-specific and age- for publication.
specific estimates for each subgroup and year,29 and
then age-standardised these estimates.28 Using the Results
bootstrap method, we drew 1000 samples, each Our analysis included 358 population-based surveys
containing 80% of all survey data, to produce uncertainty done between 2001 and 2016, with 1·9 million participants
intervals [UI] for these estimates, representing the from 168 countries (appendix p 1), representing 96% of

www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7 5

 Articles

Women Men Prevalence of insufficient physical activity ranged from

Low income Low income 16·3% (95% UI 14·3–20·7) in Oceania to 39·1%
100 (37·8–40·6) in Latin America and the Caribbean in 2016
80 (table 2). Between 2001 and 2016, the prevalence of physical
inactivity increased by more than 5 percentage points in
Prevalence (%)

60
high-income Western countries (from 30·9%, 26·4–38·1 in
40 2001, to 36·8%, 34·6–38·4, in 2016) and in Latin America
20 and Caribbean (from 33·4%, 29·1–38·6, in 2001, to 39·1%,
37·8–40·6, in 2016), whereas east and southeast Asia had a
0
decrease of more than 5 percentage points (from 25·7%,
Middle income Middle income 20·6–29·4, in 2001, to 17·3%, 15·8–22·1, in 2016).
100
Across all regions, with the exception of east and
80 southeast Asia, women were less active than men in 2016
Prevalence (%)

60
(table 2). There was a difference between sexes of more
than 10 percentage points in central Asia, Middle East
40
and north Africa; high-income Western countries; and
20 south Asia (table 2). The highest levels of insufficient
0
activity (>40%) among women in 2016 were in Latin
America and the Caribbean, south Asia, and high-income
High income High income Western countries (table 2). The lowest levels of physical
100
activity in men (<20%) in 2016 were in Oceania, east and
80 southeast Asia, and sub-Saharan Africa (table 2).
Prevalence (%)

60 The prevalence of insufficient physical activity in

high-income countries was more than double the
40
prevalence in low-income countries in 2016 (table 2). The
20 prevalence increased over time in high-income countries,
0 from 31·6% (27·1–37·2) in 2001, to 36·8% (35·0–38·0)
2000 2005 2010 2015 2000 2005 2010 2015 in 2016, whereas it was stable in low-income countries, at
Year Year
16·0% (12·0–19·6) in 2001, and 16·2% (14·2–17·9) in
Figure 3: Trends in insufficient physical activity for three income groups from 2001 to 2016 2016 (table 2, figure 3).
The shaded areas show 95% uncertainty intervals. Country-specific, age-standardised prevalence of in­
sufficient physical activity in 2016 ranged from 5·5%
the world’s population. The availability of data across (4·0–7·6) in Uganda to 67·0% (58·6–74·3) in Kuwait
income groups and regions was spread fairly evenly, with (appendix pp 28–31). In four countries, the prevalence of
the exception of Latin American and Caribbean countries insufficient physical activity was more than 50% (Kuwait,
where just more than half of countries had data (table 1). American Samoa, Saudi Arabia, and Iraq), whereas
Of the 358 surveys included, 301 were nationally the prevalence was lower than 10% in seven countries
representative, and 150 of 168 countries had at least one (Uganda, Mozambique, Lesotho, Tanzania, Niue, Vanuatu,
national survey (appendix pp 2–15). and Togo). In 55 (32·7%) of 168 countries, more than a
65 of 168 countries had at least two surveys using third of the population was insufficiently physically active.
the same questionnaire and survey coverage (table 1, In 159 of 168 countries, prevalence of insufficient
appendix p 16). Availability of data meeting these physical activity was lower in men than in women, with a
two criteria was skewed towards higher-income countries, difference of at least 10 percentage points in 65 countries,
with only 3·2% of low-income countries and 48·0% of and a difference of more than 20 percentage points in
high-income countries having trend data (table 1). nine countries: Barbados, Bahamas, Saint Lucia, Palau,
Globally, more than a quarter of adults (27·5%, 95% UI Iraq, Bangladesh, Trinidad and Tobago, Iran, and Saudi
25·0–32·2) were insufficiently physically active in 2016 Arabia. Figures 4 and 5 show country prevalence for men
(table 2). Between 2001 and 2016, levels of insufficient and women.
physical activity have decreased only marginally and Of the 65 countries with data for trends over time,
insignificantly, with a global prevalence of 28·5% 28 had decreasing levels of insufficient activity, whereas
(23·9–33·9) in 2001. Women were less active than men, levels were increasing in 37 countries. The largest
with a prevalence difference of 6 percentage points decreases (>15%) have occurred in Cook Islands, Jordan,
between sexes in 2001 (25·5%, 95% UI 23·1–33·0 for Tokelau, Samoa, Myanmar, Solomon Islands, and
men, and 31·5%, 95% UI 27·9–37·8 for women), and Tonga, while the largest increases (>15%) occurred in
of more than 8 percentage points in 2016 (23·4%, Brazil, Bulgaria, Germany, Philippines, and Singapore.
21·1–30·7, for men, and 31·7%, 28·6–39·0, for women; The average change across all of 65 countries was less
figures 1 and 2). than 0·01%.

6 www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7

 Articles

Prevalence (%)
<20·0
20·0–29·9
30·0–39·9
40·0–49·9
≥50·0
No data

Figure 4: Country prevalence of insufficient physical activity in men in 2016

Prevalence (%)
<20·0
20·0–29·9
30·0–39·9
40·0–49·9
≥50·0
No data

Figure 5: Country prevalence of insufficient physical activity in women in 2016

Asia Pacific, and prevalence was more than double in

Discussion high-income countries than in low-income countries in
Our analysis, including data from nearly 2 million 2016. In wealthier countries, the transition towards
participants (representing 96% of the global population), more sedentary occupations and personal motorised
shows that globally, in 2016, more than a quarter of all transportation probably explains the higher levels of
adults was not getting enough physical activity. This puts inactivity. Conversely, in lower-income countries, more
more than 1·4 billion adults at risk of developing or activity is undertaken at work and for transport; however,
exacerbating diseases linked to inactivity, and needs to be these behaviours are changing rapidly.31 National policy
urgently addressed. Previous estimates4 showed a global needs to be implemented to encourage non-motorised
prevalence of insufficient physical activity of 23·3% in 2010. modes of transportation, such as walking and cycling,
This difference from our results is due to our inclusion of and to promote participation in active recreation and
nearly twice as many surveys (358 vs 195), and our ability to sports in leisure time. Such policies are particularly
produce estimates for an additional 22 countries. Our important in countries with rapid urbanisation, such as
estimates will be made available in the WHO Global Health Argentina, Brazil, and Colombia,27 which contribute to
Observatory,17 where they can be downloaded for use in the high levels of insufficient activity in Latin America
projects such as burden of disease calculations.16 and the Caribbean. Effective policies include improved
Prevalence of insufficient physical activity varied provision of cycling and walking infrastructure,
greatly across regions and income groups in 2016. We improving road safety, and creating more opportunities
found the highest levels in Latin America and the Carib­ for physical activity in public open spaces and parks, in
bean, high-income Western countries, and high-income workplaces, and in other local community settings.5,32,33

www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7 7

 Articles

We also found a wide variation in country prevalence of Our study is affected by several limitations. First, similar
insufficient activity both across and within regions. The to all global analyses, data were not available for every
prevalence of insufficient activity was lower than 10% in country and year, and availability varied across countries
a few countries, and more than 50% in others. This and regions. Latin America and the Caribbean,
inequality has also been confirmed by Althoff and high-income Western countries, Oceania, and sub-
colleagues,34 who used movement sensors built into smart­ Saharan Africa had the lowest proportion of countries
phones to assess physical activity. Despite the difference in with data. However, within these regions, the most
measurement method (we used self-reported physical populous countries were more likely to have data, so that
activity), and differences in sampling, the resulting the proportion of the population with data in these regions
patterns of activity are similar, with some of the least active was still high. Countries with a population of more than
countries located in central and south America and north 10 million with no data were Bolivia, Haiti, and Peru in
Africa and the Middle East, and some of the most active Latin America and the Caribbean; Angola, Burundi,
countries in east Asia and eastern Europe. The fact that Somalia, South Sudan, and Sudan in sub-Saharan Africa;
activity varies greatly across countries, even within regions, Afghanistan in south Asia; Syria and Yemen in central
suggests that the factors that influence inactivity lie mostly Asia, Middle East, and North Africa; and North Korea in
at the national, sub­national, or community level, which is east and southeast Asia. Some of these countries are also
where policies are needed to increase physical activity.32 classified as low-income countries, which had the lowest
Our study confirms findings34–37 of lower activity in data availability of the income groups. The availability of
women than in men, with some of the biggest differences data for trends in physical inactivity was clearly skewed
in south and central Asia and the Middle East and north towards high-income countries, with 48·0% of countries
Africa. One way to explain sex differences in activity is being covered, compared with 3·2% of low-income
to assess male and female participation in different countries. In fact, our trend estimations for low-income
domains of activity (activity at work or in the house­ countries were based on one country only, Benin, which
hold, for transport, and during leisure time), and at limits the representativeness for other low-income
different intensities (moderate and vigorous). Previous countries. Data coverage for trends was low in sub-
research36,38,39 indicates that women tend to do less Saharan Africa, Latin American and the Caribbean, and
leisure-time activity, and lower-intensity activity than do south Asia, indicating that trend estimates for these
men. Offering more opportunities for safe and accessible regions should be interpreted with caution.
leisure-time activity to women in order to increase their Second, data quality also varied across countries and
overall levels of activity would therefore help close the over time. 131 (37%) of the 358 surveys included data that
gender gap and achieve the 2025 global physical activity were collected through the WHO Stepwise Approach to
target.38 Furthermore, cultural norms, traditional roles, NCD Risk Factor Surveillance, using the Global Physical
or lack of social and community support might lead to Activity Questionnaire.9,10,14 The WHO Stepwise Approach
reduced participation in physical activity among girls was started in 2001, and the earlier surveys tended to be
and women. Understanding and addressing these less representative with sometimes only subnational
barriers is needed to plan and deliver culturally sensitive coverage.14 Nonetheless, we included these data in our
actions to support behaviour change.40 analysis because they help to provide global coverage for
Our analysis was the first to assess trends in physical the earlier years.
inactivity over time. The global prevalence of physical Third, we had to rely on self-reported data in our analysis
inactivity was stable between 2001 and 2016, suggesting despite their limitations.43 Nationally repre­sentative data
no progress in reducing global levels to reach the 2025 for physical activity that are collected objectively, using
global physical activity target.6 However, we found a wide accelerometers, for example, are mostly only available for
variation in trends in inactivity across regions, income high-income countries, and results are not comparable
groups, and countries. The largest increases in insufficient across countries because of variations in data collection
physical activity have occurred in high-income countries, methods, data processing, and scoring.4,44 In the next few
whereas the largest decreases have occurred in east and years we expect that more objective physical activity data
southeast Asia. These decreases are largely explained by will become available at a larger scale, and for the next
increased participation in physical activity in China, the update of our comparable estimates we will reconsider
most populous country in the region (data not shown). inclusion of these data in some way.
Leisure-time physical activity in China has increased,41 Finally, in some cases, our estimates are different to
which might be explained by increased park use and prevalence estimates produced by countries. There are
physical activity among China’s rapidly growing elderly several reasons for this difference. We adjusted our
population.42 The lack of progress in other regions might estimates for several factors when necessary. We also
be explained by the fact that, although more than 70% of standardised our estimates to an artificial age structure,
countries have an operational physical activity policy, the the WHO Standard Population.28 Furthermore, some
scale and reach of its implementation is yet to have a countries tend to produce estimates for only leisure time
national impact.4,32 physical activity, whereas we include four domains of

8 www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7

 Articles

physical activity. In these cases, national estimates 9 WHO. Global Physical Activity Surveillance. http://www.who.int/
of physical inactivity prevalence will be higher than our ncds/surveillance/steps/GPAQ/en/ (accessed Dec 2, 2017).
10 Armstrong T, Bull F. Development of the World Health
comparable estimates. Organization Global Physical Activity Questionnaire (GPAQ).
Our data show that progress towards the global target J Public Health 2006; 14: 66–70.
set by WHO member states to reduce physical inactivity 11 WHO. WHO World Health Survey. http://www.who.int/healthinfo/
survey/en/ (accessed Dec 20, 2017).
by 10% by 2025 has been too slow and is not on track.
12 Gesis. The European Commission’s Eurobarometer Surveys.
Levels of insufficient physical activity are particularly high https://www.gesis.org/eurobarometer-data-service/home/
and still rising in high-income countries, and worldwide, (accessed Sept 30, 2017).
women are less active than are men. A significant increase 13 WHO. STEPwise approach to Surveillance (STEPS). http://www.
who.int/ncds/surveillance/steps/en/ (accessed Dec 22, 2017).
in national action is urgently needed in most countries to 14 Riley L, Guthold R, Cowan M, et al. The World Health Organization
scale-up implementation of effective policies. The Global STEPwise Approach to Noncommunicable Disease Risk-Factor
Action Plan on Physical Activity 2018–2030,32 is a new Surveillance: methods, challenges, and opportunities.
Am J Public Health 2016; 106: 74–78.
catalyst for global action, and provides a selection of
15 Hallal PC, Andersen LB, Bull FC, et al. Global physical activity
20 specific policies targeting different settings and levels: surveillance progress, pitfalls, and prospects. Lancet 2012;
populations that can be adapted and tailored to local 380: 247–57.
contexts in all countries. However, implementation will 16 Lee IM, Shiroma EJ, Lobelo F, et al. Effect of physical inactivity on
major non-communicable diseases worldwide: an analysis of
require bold leadership and full engagement across burden of disease and life expectancy. Lancet 2012; 380: 219–29.
sectors to change the current approach. Collaboration 17 WHO. Global Health Observatory (GHO) data. Prevalence of
across sectors could generate significant returns, because insufficient physical activity. http://www.who.int/gho/ncd/risk_
factors/physical_activity/en/ (accessed Oct 10, 2017).
policies that support increasing physical activity can 18 The World Bank. List of economies, December 2016. databank.
provide other benefits to health, local economies, worldbank.org/data/download/site-content/CLASS.xls
community wellbeing, and environ­mental sustainability, (accessed May 13, 2017).
19 NCD Risk Factor Collaboration. Worldwide trends in body-mass
and contri­ bute towards achieving many of the 2030 index, underweight, overweight, and obesity from 1975 to 2016:
Sustainable Development Goals.5­ a pooled analysis of 2416 population-based measurement studies in
128·9 million children, adolescents, and adults. Lancet 2017;
Contributors
390: 2627–42.
RG analysed the data and wrote the first draft of the report. GAS and RG
20 NCD Risk Factor Collaboration. Worldwide trends in blood pressure
developed the methodological approach, with inputs from all other
from 1975 to 2015: a pooled analysis of 1479 population-based
authors. All authors designed the study, and revised and approved the
measurement studies with 19·1 million participants. Lancet 2017;
final report. 389: 37–55.
Declaration of interests 21 Pate RR, Pratt M, Blair SN, et al. Physical activity and public health.
All authors are staff members of WHO. The authors alone are A recommendation from the Centers for Disease Control and
responsible for the views expressed in this publication and they do not Prevention and the American College of Sports Medicine. JAMA
necessarily represent the decisions, policy, or views of WHO. 1995; 273: 402–07.
22 WHO. Assessing national capacity for the prevention and control of
Acknowledgments NCDs. http://www.who.int/ncds/surveillance/ncd-capacity/en/
This study received no funding. See appendix (p 32) for further (accessed July 10, 2017).
acknowledgments. 23 Rzewnicki R, Vanden Auweele Y, De Bourdeaudhuij I.
Addressing overreporting on the International Physical Activity
References
Questionnaire (IPAQ) telephone survey with a population sample.
1 Physical Activity Guidelines Advisory Committee report, 2008.
Public Health Nutr 2003; 6: 299–305.
To the Secretary of Health and Human Services. Part A: executive
summary. Nutr Rev 2009; 67: 114–20. 24 Hallal PC, Gomez LF, Parra DC, et al. Lessons learned after 10 years
of IPAQ use in Brazil and Colombia. J Phys Act Health 2010;
2 Warburton DE, Charlesworth S, Ivey A, Nettlefold L, Bredin SS.
7 (suppl 2): S259–64.
A systematic review of the evidence for Canada’s Physical Activity
Guidelines for Adults. Int J Behav Nutr Phys Act 2010; 7: 39. 25 Ekelund U, Sepp H, Brage S, et al. Criterion-related validity of the
last 7-day, short form of the International Physical Activity
3 WHO. Global recommendations on physical activity for health.
Questionnaire in Swedish adults. Public Health Nutr 2006;
Geneva: World Health Organization, 2010.
9: 258–65.
4 Sallis JF, Bull F, Guthold R, et al. Progress in physical activity over
26 Ainsworth BE, Macera CA, Jones DA, et al. Comparison of the 2001
the Olympic quadrennium. Lancet 2016; 388: 1325–36.
BRFSS and the IPAQ Physical Activity Questionnaires.
5 ISPAH International Society for Physical Activity and Health. Med Sci Sports Exerc 2006; 38: 1584–92.
The Bangkok Declaration on Physical Activity for Global Health
27 United Nations. Department of Economic and Social Affairs.
and Sustainable Development. Br J Sports Med 2017; 51: 1389–91.
Population Division. World Urbanization Prospects, the 2014
6 WHO. Global action plan for the prevention and control of revision. https://esa.un.org/unpd/wup/ (accessed Oct 13, 2017).
noncommunicable diseases 2013–2020. Geneva: World Health
28 Ahmad O, Boschi-Pinto C, Lopez AD, Murray CJL, Lozano R,
Organization, 2013.
Inoue M. Age standardization of rates: a new WHO standard.
7 Bull FC, Armstrong TP, Dixon T, Ham S, Neiman A, Pratt M. Geneva: World Health Organization; 2001.
Physical Inactivity. In: Ezzati M, Lopez A, Rodgers A, Murray CJL,
29 United Nations. Department of Economic and Social Affairs.
eds. Comparative quantification of health risks. Global and regional
Population Division. Monitoring Global Population Trends.
burden of disease attributable to selected major risk factors.
http://www.un.org/en/development/desa/population/publications/
Geneva: World Health Organization; 2004.
index.shtml (accessed Oct 13, 2017).
8 International Physical Activity Questionnaire. Guidelines for Data
30 Stevens GA, White RA, Flaxman SR, et al. Global prevalence of
Processing and Analysis of the International Physical Activity
vision impairment and blindness: magnitude and temporal trends,
Questionnaire (IPAQ). 2005. https://www.researchgate.net
1990–2010. Ophthalmology 2013; 120: 2377–84.
/file.PostFileLoader.html?id=5641f4c36143250eac8b45b7&assetKey
=AS%3A294237418606593%401447163075131 (accessed 31 Ng SW, Popkin BM. Time use and physical activity: a shift away
July 5, 2017). from movement across the globe. Obes Rev 2012; 13: 659–80.

www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7 9

 Articles

32 WHO. Global action plan on physical activity 2018–2030. More active 39 Guthold R, Louazani SA, Riley LM, et al. Physical activity in
people for a healthier world. Geneva: World Health Organization; 22 African countries: results from the World Health Organization
2018. STEPwise approach to chronic disease risk factor surveillance.
33 Stevenson M, Thompson J, de Sá TH, et al. Land use, transport, and Am J Prev Med 2011; 41: 52–60.
population health: estimating the health benefits of compact cities. 40 WHO Regional Office for the Eastern Mediterranean. Promoting
Lancet 2016; 388: 2925–35. physical activity in the Eastern Mediterranean Region through a
34 Althoff T, Sosič R, Hicks JL, King AC, Delp SL, Leskovec J. life-course approach. Cairo: World Health Organization; 2014.
Large-scale physical activity data reveal worldwide activity inequality. 41 Tian Y, Jiang C, Wang M, et al. BMI, leisure-time physical activity,
Nature 2017; 547: 336–39. and physical fitness in adults in China: results from a series of
35 Ranasinghe CD, Ranasinghe P, Jayawardena R, Misra A. national surveys, 2000–14. Lancet Diabetes Endocrinol 2016;
Physical activity patterns among south-Asian adults: a systematic 4: 487–97.
review. Int J Behav Nutr Phys Act 2013; 10: 116. 42 Tu H, Liao X, Schuller K, et al. Insights from an observational
36 Koohpayehzadeh J, Etemad K, Abbasi M, et al. Gender-specific assessment of park-based physical activity in Nanchang, China.
changes in physical activity pattern in Iran: national surveillance of Prev Med 2015; 2: 930–34.
risk factors of non-communicable diseases (2007–2011). 43 Ara I, Aparicio-Ugarriza R, Morales-Barco D,
Int J Public Health 2014; 59: 231–41. Nascimento de Souza W, Mata E, González-Gross M. Physical
37 Global Observatory for Physical Activity (GoPA). Country Cards. activity assessment in the general population; validated self-report
http://www.globalphysicalactivityobservatory.com/country-cards/ methods. Nutr Hosp 2015; 31 (suppl 3): 211–18.
(accessed Jan 2, 2018). 44 Wijndaele K, Westgate K, Stephens SK, et al. Utilization and
38 Mielke GI, da Silva ICM, Kolbe-Alexander TL, Brown WJ. harmonization of adult accelerometry data: review and expert
Shifting the physical inactivity curve worldwide by closing the consensus. Med Sci Sports Exerc 2015; 47: 2129–39.
gender gap. Sports Med 2018; 48: 481–89.

10 www.thelancet.com/lancetgh Published online September 4, 2018 http://dx.doi.org/10.1016/S2214-109X(18)30357-7