Physical Inactivity
![Octet Stream icon](https://www.nutrans.org/china/application.png)
Physical inactivity is recognized as an important risk factor for multiple causes of death and chronic morbidity and disability. Physical activity was chosen rather than physical fitness as the measure of exposure because it is through increases in the behaviour (physical activity) that health benefits accrue and improvements in cardiorespiratory fitness can be achieved. Moreover, there were insufficient data available worldwide to consider fitness as the exposure. Exposure was assessed as a trichotomous variable to avoid limiting the assessment of total burden to only that associated with the highest risk, namely the most inactive (a dichotomous approach). However due to a lack of data on physical inactivity, use of a more detailed (continuous) exposure variable was not possible nor was the use of a fourth category of “high activity”. Therefore, our estimates of burden are likely to underestimate the total attributable burden to inactivity because of limitations with measures of exposure. Level 1 exposure (inactive) was defined as “doing no or very little physical activity at work, at home, for transport or in discretionary time”. Level 2 exposure (insufficiently active) was defined as“doing some physical activity but less than 150 minutes of moderate-intensity physical activity or 60 minutes of vigorous-intensity physical activity a week accumulated across work, home, transport or discretionary domains”. We found a wide range of survey instruments and methodologies have been used for collecting, analysing and reporting data on physical activity. Most data were available for discretionary-time activity, some data were found on occupational activity and little and no national data were available for transport- and domestic-related activity, respectively. A comprehensive literature search and contact with key agencies and known researchers uncovered over 50 data sets on physical inactivity in adult populations covering 43 countries across 13 subregions. However, only 21 data sets covering 32 countries met our inclusion criteria. Hierarchical modelling techniques were used to predict discretionary-time activity using age, sex, geographic region and a measure of tertiary education. Linear regression was used to predict occupational activity and transport-related activity using two World Bank indicators (% employed in agriculture and car ownership, respectively). We used these estimates to compute the level of total inactivity for 145 countries and aggregated these data to create estimates for 14 subregions. The final global estimate for total inactivity (level 1 exposure) was 17.1% and this ranged from 10.3% in AFR-D to 24.8% in EUR-C. Across most but not all subregions females were slightly more inactive than males and younger adults were less inactive than older adults (range 9.6–46.8% across the 12 age-sex categories). The final global estimate for insufficient activity (level 2 exposure) was 40.6% and this ranged from 31.7% in AMR-D to 51.5% in WPR-A. The independent causal relationship between physical inactivity and ischaemic heart disease, ischaemic stroke, type II diabetes, colon cancer and breast cancer is well established; we provided new estimates of the magnitude of risk associated with inactivity. A comprehensive search of literature from 1980 onwards identified well over 100 studies assessing the relationship between physical inactivity and the set of health outcomes that met our criteria. Also, several quantitative and qualitative reviews of the association between physical inactivity and ischaemic heart disease and stoke were found but there were no quantitative meta-analyses for breast cancer, colon cancer and type II diabetes. Most of the epidemiological studies meeting our inclusion criteria measured discretionary-time activity, some studies assessed occupational activity but only a few studies incorporated transport-related activity. No study included domestic-related physical activity. Given these data and differences between previous work and our definition of exposure, we completed a series of new meta-analyses for each health outcome. To address concerns regarding measurement error associated with physical activity, an adjustment factor was incorporated into the meta-analyses. All risk estimates were attenuated for ages 70 and over. There is emerging consensus on the protective effects of activity in regards to preventing falls, osteoarthritis and osteoporosis and impaired mental health but these disease end-points did not meet our inclusion criteria. Globally physical inactivity accounted for 21.5% of ischaemic heart disease, 11% of ischaemic stroke, 14% of diabetes, 16% of colon cancer and 10% of breast cancer. The results show small differences between males and females, due in part to differences in level of exposure and to different distribution of events between men and women. In summary, physically inactive lifestyles accounted for 3.3% of deaths and 19 million disability-adjusted life years (DALYs) worldwide. There were small, non-significant differences in the attributable fractions across subregions. Due to our conservative methods and a number of important limitations, our global estimates are likely to be an underestimate of the true burden attributable to inactive lifestyles.
CHAP
Comparative quantification of health risks: Global and regional burden of disease attributable to selected major risk factors
Bull, Fiona C.
Armstrong, Timothy P.
Dixon, Tracy
Ham, Sandra
Neiman, Andrea
Pratt, Michael
Ezzati, Majid
Lopez, Alan D.
Rodgers, Anthony
Murray, Christopher J. L.
2004
1
729-881
World Health Organization
Geneva
326