1. Home
  2. ENLACE: Data Portal on Noncommunicable Diseases, Mental Health, and External Causes
  3. ENLACE data portal - Technical Notes

ENLACE data portal - Technical Notes

ENLACE TECHNICAL NOTES ABOUT

 

Background

This document presents methodological information on the health measures used throughout several data products published in the Noncommunicable Disease and Mental Health Data Portal. It includes 1) mortality and burden of diseases measures from noncommunicable diseases (NCD), mental health conditions and substance use disorders, and injuries; 2) prevalence of risk factors for noncommunicable diseases.  

The data presented are derived from several sources, each of which is explained in the following notes.

Demographic indicators

The estimates of the population size by age, sex, year, and country were obtained from the World Population Prospects, Revision 2019. Further details on the estimation methods are published in the World Population Prospects Report 2019 (1).

Mortality and burden of disease outcome indicators

Measures of mortality and burden of diseases such as deaths, probability of dying,  disability-adjusted life years, years lived with disability, and years of life lost due to premature deaths are presented in the NMH Data Portal. 

Measures

  • Deaths
  • The unconditional probability of dying between exact ages 30 to 70 from any of the four major noncommunicable diseases (cardiovascular diseases, cancers, diabetes mellitus, and chronic respiratory disease). 
  • Disability-adjusted Life Years (DALYs) 
  • Years of Life Lost due to Premature Mortality
  • Years Lived with Disability (YLDs)

Disability-Adjusted Life Years (DALYs)
The overall burden of disease is assessed using the disability-adjusted life year (DALY), a time-based measure that combines years of life lost due to premature mortality (YLLs) and years of life lost due to time lived in states of less than full health, or years of healthy life lost due to disability (YLDs). One DALY represents the loss of the equivalent of one year of full health. Using DALYs, the burden of diseases that cause premature death but little disability (such as drowning or measles) can be compared to that of diseases that do not cause death but do cause disability (such as cataracts causing blindness). DALYs are the sum of YLLs and YLDs. Detailed method of estimation is available at: https://cdn.who.int/media/docs/default-source/gho-documents/global-heal…;

Years of Life Lost due to premature mortality (YLLs)
Years of life lost (YLL) is a time-based measure of premature mortality that takes into account both the frequency of deaths and the age at which it occurs. One YLL represents the loss of one year of life. At the individual level, YLLs are calculated by subtracting the age at death from the longest possible life expectancy for a person at that age. At the population level, YLLs are calculated from the number of deaths multiplied by the standard life expectancy (SLE) at the age at which death occurs. 

The YLLs for a cause are essentially calculated as the number of cause-specific deaths multiplied by a loss function specifying the years lost for deaths as a function of the age at which death occurs. The basic formula for YLLs is the following for a given cause c, age a, sex s, and year t: 
YLL(c,s,a,t) = N(c,s,a,t) x L(s,a) 
where: N(c,s,a,t) is the number of deaths due to the cause c for the given age a and sex s in year t. L(s,a) is a standard loss function specifying years of life lost for a death at age a for sex s. The number of deaths is obtained from the WHO Global Health Estimates, and the standard loss function is based on the frontier national life expectancy projected for the year 2050 by the World Population Prospects 2012 (UN Population Division, 2013), with a life expectancy at birth of 92 years. Detailed method of estimation is available at: https://cdn.who.int/media/docs/default-source/gho-documents/global-heal…;

Years Lived with Disability (YLDs)
YLD is a time-based measure of healthy life lost due to disability or ill health. One YLD represents the equivalent of one full year of healthy life lost due to disability or living in less-than-ideal health. YLDs are estimated by multiplying the prevalence counts with the disability weight for a given disease, condition, or injury. Disability weights represent the severity or the magnitude of health loss associated with specific health outcomes. Detailed method of estimation is available at: https://cdn.who.int/media/docs/default-source/gho-documents/global-heal…
 

Metrics

Measures of deaths, DALYs, YLLs, and YLDs are presented in absolute numbers and rates per population by age, sex, cause, year, and location, having the metric deaths for mortality measures, and years for the burden of disease measures. The probability of dying is presented in percentages.  

Causes of the burden of diseases 

Data are presented for categories and subcategories of noncommunicable diseases according to the Global Health Estimates list of causes. The categories of fatal and non-fatal diseases and conditions in the GHE list are arranged in a hierarchy of four levels of aggregation, where the first level includes three broad groups: 1) Communicable, maternal, neonatal, and nutritional conditions; 2) Noncommunicable Diseases, and 3) Injuries. Within each of those categories, causes are broken down with increasing specificity at each level. The list is mutually exclusive and collectively exhaustive at every level of aggregation, and those diseases not individually specified are captured in residual categories.

The full list of cause categories and corresponding ICD-10 codes is detailed in the technical document WHO methods and data sources for country-level causes of death 2000-2019, Annex A, page 62.

Method of estimation

Noncommunicable disease mortality and burden of disease indicators were calculated based on the estimated number of deaths, DALYs, YLLs, and YLDs from the WHO Global Health Estimates (GHE) 2000-2019 (2). 

Estimated deaths, DALYs, YLLs, and YLDs by age, sex, cause of death, and year for the Member States of the Pan American Health Organization/World Health Organization (35 Member States, excluding Dominica, and Saint Kitts and Nevis because they have a population below 90,000 inhabitants in 2019) were extracted from the WHO GHE 2000-2019 dataset (2). The WHO methods and data sources for the GHE estimates 2000-2019 are extensively documented elsewhere (3-4). In summary, data from national civil registration and vital statistics (CRVS) and/or mortality information systems reported to PAHO and WHO by national authorities were used as the main data source. Mortality data was corrected by missing sex and age, and deaths were rescaled by sub-registration level. Cause-of-death data quality issues due to diagnostic and coding accuracy were adjusted using formal demographic techniques called death distribution methods (DDM). For instance, deaths with underlying causes of death coded to ill-defined and garbage codes are redistributed to well-defined causes and mapped to the WHO Global Health Estimates (GHE) cause list (3-4).

Noncommunicable diseases (NCD) comprise all disease categories and subcategories from Group II of the GHE list of causes. The four major NCD (4NCD) deaths are those with underlying cause-of-death coded as cardiovascular diseases (I00-I99), cancer (C00-C97), diabetes (E10-E14), and chronic respiratory diseases (J30-J98), according to the GHE cause-of-death list (3-4).

Age-specific rates and crude rates by sex, location, and year for each cause were calculated by dividing the number of deaths, DALYs, YLLs, or YLDs by the population size and multiplying it by 100,000 using the World Population Prospects, 2019 Revision database. (5) Age-standardized death rates by country and year were calculated by the direct method of rate standardization using the WHO World Standard Population. (7) 

Rates

Three types of rates are presented in the NMH Data Portal's information products: crude rates, age-specific rates, and age-standardized rates.

Crude and age-specific rates are equal to the total number of deaths, DALYs, YLLs, YLDs in a specific year in the population category of interest, divided by the at-risk population for that category and multiplied by 100,000.

Crude rates don't take into account the age distribution of the population, so they are influenced by the underlying age distribution of the country’s population. In one geographic area or population group with an older population, the death rate will be higher than a location with a younger population as mortality increases with age. For this reason, crude rates are not a good measure for comparing mortality between two or more countries, geographic areas, or population groups.

As the age distribution of a population (the number of people in particular age categories) can change over time and can be different among geographic areas and countries, it is recommended to use age-standardized rates as an unbias measure when we want to compare the outcome measure across population groups or analyze their trends over time. 

Age-standardized rates ensure that variations in the outcome measure across geographic areas or periods of time are not due to differences in the age distribution of the populations being compared. The age-standardized rates presented in the NMH Data Portal have been computed by the method of direct standardization of rates using the World Standard Population. 

The direct method of rate standardization

An age-standardized rate is a weighted average of the age-specific (crude) rates, where the weights are the proportions of persons in the corresponding age groups of a standard population. The potential confounding effect of age is reduced when comparing age-standardized rates computed using the same standard population.

The unconditional probability of dying from NCDs

The risk of premature death from target NCDs was measured using the unconditional probability of dying between exact ages 30 and 70 years from any of the four major NCDs. It refers to the probability of dying without any competing cause of death, and methods for estimation are based on life tables informed by age-specific death rates.

The unconditional probability of dying between exact ages of 30 to 70 years was estimated for the period 2000-2019 using age-specific death rates (in 5-year age groups, e.g. 30-34, 35-39, … 60-64, 65-69) from any of the four major NCDs (cardiovascular diseases (I00-I99), cancer (C00-C97), diabetes (E10-E14) and chronic respiratory diseases (J30-J98), for each Member State (2). Using the life table method, the risk of death between the exact ages of 30 and 70 from any of the four NCDs in the absence of other competing causes of death, was calculated as described below.

Five-year death rates were calculated using the equation below:

                     5Mx = Total deaths from the four NCDs between exact age x and x+5 / Total population between exact age x and x+5

Five-year death rates were then translated into the probability of death from the four major NCDs in each five-year age range using the following equation:

                    5qx = ( 5Mx * 5 ) / 1 + 5Mx * 2.5 

And the unconditional probability of death for the 30 to 70 age range, was calculated using the equation:

                    40q30 = 1 - { (1- 5q30 ) x ... x (1- 5q65 ) }

Detailed information about the unconditional probability of dying between exacta gest of 30 to 70 years from any of the four major NCDs is described in the NCD Global Monitoring Framework: Indicator Definitions and Specifications [pdf file, 465 KB].

NCD Risk Factor Indicators

Prevalence estimates are given for the behavioral, metabolic, and environmental risk factors defined below:

  • Total alcohol per capita consumption (APC), in liters of pure alcohol: total (sum of recorded APC and unrecorded APC) amount of alcohol consumed per person (15 years and older) over a calendar year, adjusted for tourist consumption, in liters of pure alcohol.
  • Insufficient physical activity: the percentage of the population aged 18 years and older who were physically inactive – defined as not meeting the WHO recommendations on physical activity for health: 150 minutes of moderate-intensity physical activity per week or 75 minutes of vigorous-intensity physical activity per week or an equivalent combination of moderate- and vigorous-intensity physical activity.
  • Salt intake: the mean population salt intake in grams per day among adults aged 25 years and older.
  • Current tobacco smoking: the percentage of the population aged 15 years and older who smoke any tobacco products.
  • Current tobacco use in adolescents: the percentage of the youth population aged 13-15 years who used some smoked or smokeless tobacco product at least once in the 30 days prior to the survey. 
  • Raised blood pressure: the percentage of the population aged 18 years and older having systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥90 mmHg.
  • Raised blood glucose: the percentage of the population aged 18 years and older who have fasting plasma glucose of 7.0mmol/L or higher, or a history of diagnosis with diabetes, or use of insulin or oral hypoglycemic drugs.
  • Obesity:
    • Adults: the percentage of the population aged 18 years and older having a body mass index (BMI) ≥30 Kg/m2
    • Adolescents: the percentage of the population aged 10–19 years who are more than 2 SD above the median of the WHO growth reference for children and adolescents.
  • Ambient air pollution: the exceedance of the WHO guideline level for the annual mean concentration of particles of ≤ 2.5 micrometers in the air (proportion).
  • Household air pollution: the percentage of the population with primary reliance on polluting fuels and technologies.

Methods of estimation

The primary data source for the estimates of total alcohol per capita consumption (APC) was government data on recorded alcohol per capita consumption supplied by the respective Member States. If these data were not available, data from economic operators and the Food and Agriculture Organization of the United Nations statistical database (FAOSTAT) were used. The total per capita consumption of alcohol in 2016 was calculated from a three-year average of recorded (for 2015, 2016, and 2017) per capita consumption and applying unrecorded proportion (for 2016) and tourist consumption (for 2016) of tourists visiting the country and inhabitants visiting other countries. For male and female per capita consumption, the proportion of alcohol consumed by men versus women, and the UN Population Division population estimates for 2016 (4), were used. Further details on the estimation methods are published in the Global Status Report on Alcohol and Health 2018 (8).

For the adult insufficient physical activity crude-adjusted estimates, data were pooled from population-based surveys reporting on insufficient physical activity prevalence, which included self-reported physical activity at work, at home, for transport, and during leisure time. Regression models were used to adjust survey data to a standard definition and standard age groups. In order to derive a standard year, time trends were estimated using multilevel mixed-effects modeling. Full methodological details have been published (9).

Age-standardized estimates for sodium intake (grams per day) were estimated using hierarchical Bayesian estimation models based on available data from urine-based and diet-based national and regional surveys. The full methodology has been published (10). The sodium intake estimates were then converted to salt intake estimates by multiplying by 2.54.

Crude-adjusted prevalence for current tobacco smoking was estimated from national surveys that met the following criteria: i) that the survey provided national summary data for one or more of four tobacco use definitions – daily tobacco smoker, current tobacco smoker, daily cigarette smoker, or current cigarette smoker; ii) that the survey included randomly selected participants who were representative of the national population; and iii) that the survey presented prevalence rates by age and sex. Countries with no surveys, or insufficient surveys (e.g. only one survey in total, or no survey during the previous 10 years), were excluded from the analysis. Regression models were run at the UN subregional level to obtain age-and-sex-specific prevalence rates for current tobacco smoking for the years 2010–2025 (11).

Crude and age-standardized estimates of prevalence for raised blood pressure, raised blood glucose, and obesity, are based on aggregated data provided by countries to WHO or obtained through a review of published and unpublished literature. The inclusion criteria for estimation analysis stipulated that data had to come from a random sample of the general population, with clearly indicated survey methods and risk factor definitions. Detailed estimation methods have been published (12, 13, 14).

The indicator of exposure to outdoor air pollution was estimated by dividing the annual mean concentration of fine particulate matter (particles with diameters ≤ 2.5 micrometers) (PM2.5) in a country by the recommended annual mean concentration level of PM2.5 found in WHO Air Quality Guidelines: Global Update 2005 (15). Country-level estimates of PM2.5 were derived using a mathematical model that used ground-level measurements of PM compiled in the WHO outdoor air pollution database (16), data from satellite remote sensing, and other demographic data (17).

The proportion of the population in a country relying mainly on polluting fuels and technologies for cooking was used as a proxy indicator for estimating population exposure to household air pollution. Current households using mainly coal, wood, charcoal, dung, crop residues, and kerosene are considered exposed. Information on the types of fuels and technologies used by households for cooking has been regularly reported in household surveys or censuses and compiled in the WHO Household Energy Database (18). The data were further modeled to derive point estimates by county, year, and area of residence (urban/rural) (18).

Data Classification Method

Maps and bar charts in data visualizations across the NMH Data Portal presents data classified in discrete classes or ranges to facilitate data reading and interpretation. Data is classified using the quantile data classification method into five classes or quintiles. This method creates classes in a way that each one contains an equal number of observations. As we are creating quintiles, each class contains 20% of observations. The resulting quintile classes are identified as Q1, Q2, Q3, Q4, and Q5, where Q1 is the first quintile which contains the bottom 20% of observations with the lowest values, and Q5 includes the top 20% of observations with the highest values. 

Quintiles are color codes using a diverging color scheme where usually Q1 (the lowest quintile) is coded with dark blue, Q2 is coded with light blue, Q3 with yellow, Q4 is coded with orange, and Q5 (the upper quintile) is color-coded with red. 

Trends over time

For assessing the time trend, the percentage of change between the indicator's value at a given year and its value at the year 2000, was used. 

The percentage change was calculated using the following equation:

             Percentage change (%) = [ Value (current year) - Value (year 2000) ] / Value (year 2000) * 100%

The percentage change can be interpreted as follow: when it is negative implies a decreasing trend, when it is positive means an increasing trend, and when its value is zero or close to zero implies a constant trend.

The percentage change value is presented through line charts from the trends data visualizations usually together with the indicator value in the year 2016, which is the latest value of the series of estimates.

Rehabilitation

Estimates of needs for rehabilitation services 

Regarding rehabilitation, data from the global estimates of the need for rehabilitation, jointly produced by WHO and the Institute of Health Metrics and Evaluation (IHME), were used. The data sources and methods for measuring the need for rehabilitation were documented in the paper “Global estimates of the need for rehabilitation based on the Global Burden of Disease study 2019” published in the medical journal The Lancet. The paper provides the first-ever global, regional, and national figures on the number of people in need of rehabilitation. 

In brief, to estimate the need for rehabilitation, data from the Global Burden of Diseases, Injuries, and Risk Factors Study 2019 were used to calculate the prevalence and years of life lived with disability (YLDs) of 25 diseases, impairments, or aggregations of sequelae that were selected as amenable to rehabilitation. All analyses were done at the country level and then aggregated to seven regions: World Bank high-income countries and the six WHO regions (ie, Africa, the Americas, Southeast Asia, Europe, Eastern Mediterranean, and Western Pacific).

The GBD 2019 study estimated incidence, prevalence, and YLDs by age, sex, year, and location for 354 diseases and injuries, and 3484 sequelae (ie, the disabling consequences of these diseases and injuries). Prevalence accounts for the number of people (new and old cases) with a disease or health condition. YLDs are a measure of the burden of non-fatal disease and injury measured in number of years and were calculated by multiplying the prevalence of each sequela by the estimated level of health loss in the form of a disability weight. Disability weights range from 0 (ie, perfect health) to 1 (ie, death) and represent the severity of the disease. These weights were derived from population surveys using pairwise comparison methods between random pairs of health states. The disability weights were defined, measured, and given numerical value to quantify the time lived in non-fatal health states. 

Selection of conditions amenable to rehabilitation

For the selection of health conditions, a stepwise approach was followed. First, the 20 conditions with the highest number of associated YLDs were identified. Second, from these, conditions for which rehabilitation is not essential and is usually indicated as a secondary intervention (eg, dietary iron deficiency or oral disorders) were excluded. Lastly, a group of experts in the field of rehabilitation was convened by WHO to discuss the current list and add any health conditions for which rehabilitation is a key intervention as part of an overall management plan. 25 health conditions were selected for the analysis.

The 25 selected conditions were grouped and presented into seven aggregate disease and injury categories, following the GBD standard categorization of diseases. (Table of conditions amenable to rehabilitation)

Outcome measures

Prevalence and years lived with disabilities (YLDs) were estimated for all conditions, the seven aggregate categories, and individual conditions amenable to rehabilitation. Both measures we calculated by age, sex, location and year from 1990 to 2019.

Estimates of each outcome measure were corrected for comorbidity using simulation methods and assumed a multiplicative model for coexisting health states. For sequelae that fell into multiple health conditions, those cases were included in each most-detailed category but only once in the parent category. For example, a person with the sequela “autism spectrum disorder with moderate developmental intellectual disability” would be counted in both “autism spectrum disorders” and “developmental intellectual disability” but only once in “mental disorders” and in the total rehabilitation prevalence count. Similarly, a patient with heart failure due to COPD was counted in both “heart failure” and “chronic obstructive pulmonary disease” among health conditions, and “cardiovascular diseases” and “chronic respiratory diseases” among disease areas. However, when we estimated the number of people who could potentially benefit from rehabilitation, we counted this example as a single person.

Rehabilitation workforce

Workforce data and information for this topic were collected by the PAHO Regional Rehabilitation program via desk research, examination of country site visit reports, direct engagement with rehabilitation professionals in specific countries, and discussions with the ministry of health representatives through PAHO country offices.

Publicly available data on the numbers of physical and rehabilitation medicine doctors (physiatrists), physiotherapists, occupational therapists, speech and language therapists, prosthetists and orthotists, and psychologists per country were collected. Publicly available information from international professional associations such as World Physiotherapy (formerly World Confederation of Physical Therapy) and the World Federation of Occupational Therapy (WFOT)​ was used as primary data as these organizations are in close contact with leaders from the national professional associations. Data from some countries were also provided by the International Society of Physical and Rehabilitation Medicine (ISPRM), the International Society of Prosthetics and Orthotics (ISPO), or local professional associations. Data on the number of psychologists in each country were obtained from the World Mental Health Atlas which, in some countries, also contained information for occupational therapists and speech and language therapists. For countries that had undertaken a recent national rehabilitation assessment (Bolivia, El Salvador, Guyana, and Haiti) data were used from this source. The WHO Global Health Observatory was also used as a source for physiotherapists. Finally, when available, public or non-public data from Statistical Departments of Ministries of Health, or Ministries of Labor were used, especially in countries and territories without professional associations. 

Where significant discrepancies in data occurred from different sources, follow-up was made to verify the information with the Ministry of Health. If not possible to verify, a decision was taken on which source to use by the PAHO advisor on disability and rehabilitation. A master database was created detailing sources of data, which will be maintained and updated by PAHO as new data emerges. 

Workforce indicators
Two main indicators are featured in the rehabilitation workforce:

  1. Rehabilitation workforce density per 10,000 population. This indicator informs on the capacity of the health system to deliver rehabilitation to the population. It is presented as the overall rehabilitation workforce and by profession.
  2. Rehabilitation workforce per population in need of rehabilitation (per 10,000 population). This is a more sensitive and specific type of rehabilitation workforce density indicator as it accounts for the capacity of the health system to deliver rehabilitation care to those who need it.   

Data and information for these indicators were presented at the national level, and the countries and territories were clustered into four subregions:  

  • North America (Canada and the United States). 
  • Central America, Mexico, and the Spanish-speaking Caribbean.  
  • English, French, and Dutch Speaking Caribbean.  
    • English, French, and Dutch Speaking Caribbean (PAHO/WHO Member States only).
    • English, French, and Dutch Speaking Caribbean (Territories only).
  • South America.  

List of countries in the Region of the Americas and in each subregion

This section specifies the list of 35 PAHO/WHO member states of the Region of the Americas and the subregion where they are included. 

Region of the Américas: Argentina, Antigua and Barbuda, Bahamas, Barbados, Belize, Brazil, Bolivia (Plurinational State of), Canada, Chile, Colombia, Costa Rica, Cuba, Dominica, Dominican Republic, Ecuador, El Salvador, Grenada, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Suriname, Trinidad and Tobago, Paraguay, Peru, Venezuela (Bolivarian Republic of), United States of America, Uruguay.

North America: Canada, United States of America.

Central America, Mexico, and the Latin Caribbean: Costa Rica, Cuba, Dominican Republic, Guatemala, Haiti, Honduras, El Salvador, Mexico, Nicaragua, and Panama.

Non-Latin Caribbean: Antigua and Barbuda, Bahamas, Barbados, Belize, Grenada, Guyana, Jamaica, Saint Lucia, Saint Vincent and the Grenadines, Suriname, Trinidad and Tobago.

Andean Area: Bolivia (Plurinational State of), Colombia, Ecuador, Peru, Venezuela (the Bolivarian Republic of).

Southern Cone and Brazil: Argentina, Brazil, Chile, Paraguay, Uruguay.

Notes:

The World Health Organization produces global health estimates (GHE) for the WHO Member States with more than 90,000 population in 2019. Two out of 35 PAHO/WHO Member States (Dominica, and Saint Kitts and Nevis) are in this category, so no data on mortality and disease burden are available for them.

References

  1. United Nations, Department of Economic and Social Affairs, Population Division (2019). World Population Prospects: The 2019 Revision, Methodology of the United Nations Population Estimates and Projections, Technical document: ST/ESA/SER.A/425. New York: United Nations. Available online.  (accessed 1 February 2021)
  2. WHO Global Health Estimates 2000-2019. Geneva, World Health Organization; 2020. [Internet] Available online. (accessed 1 February 2021).
  3. WHO methods and data sources for country-level causes of death 2000-2019. Geneva: World Health Organization; 2020. Available online (accessed 1 February 2021).
  4. WHO methods and data sources for global burden of disease estimates, 2000-2019. Geneva: World Health Organization, 2020. Available online (accessed 1 February 2021).  
  5. UN. Department of Economic and Social Affairs. World Population Prospects – 2019 revision. New York: United Nations Population Division; 2019. (https://esa.un.org/unpd/wpp/, accessed 1 February 2021). 
  6. US Census Bureau. International Data Base (IDB). August 2017. https://www.census.gov/data-tools/demo/idb/informationGateway.php (accessed 21 May 2018)
  7. Ahmad OB, Boschi-Pinto C, Lopez AD, et al. Age standardization of rates: a new WHO standard–GPE Discussion Paper Series: No. 31. World Health Organization 2001. (pdf file: http://www.who.int/healthinfo/paper31.pdf, accessed 21 May 2018)
  8. Global status report on alcohol and health 2018. Geneva: World Health Organization; 2018. 
  9. Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufficient physical activity from 2001 to 2016: A pooled analysis of 358 population-based surveys with 1.9 million participants. Lancet Global Health. 2018. DOI: 10.1016/S2214-109X(18)30357-7. 
  10. Powles J, Fahimi S, Micha R, et al. Global, regional and national sodium intakes in 1990 and 2010: a systematic analysis of 24h urinary sodium excretion and dietary surveys worldwide. British Medical Journal. 2013;3(12):e003733. doi:10.1136/bmjopen-2013-003733. 
  11. WHO global report on trends in prevalence of tobacco smoking 2000–2025, second edition. Geneva: World Health Organization; 2018. (http://www.who.int/tobacco/publications/surveillance/trends-tobacco-smoking-second-edition/en/, accessed 16 August 2018).
  12. NCD-RisC. Worldwide trends in blood pressure from 1975 to 2015: a pooled analysis of 1479 population-based measurement studies with 19.1 million participants. Lancet. 2016;389(10064):37–55. doi:10.1016/S0140-6736(16)31919-5. 
  13. NCD-RisC. Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4.4 million participants. 2016. The Lancet 387 (10027), 1513–1530. 
  14. NCD-RisC. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies with 128.9 million participants. Lancet 2017. 
  15. WHO Air Quality Guidelines: Global update 2005. Copenhagen: World Health Organization Regional Office for Europe; 2006. (http://www.who.int/phe/health_topics/outdoorair/outdoorair_aqg/en/, accessed 17 August 2018). 
  16. WHO Global Ambient Air Quality Database. Geneva: World Health Organization. (http://www.who.int/airpollution/data/cities/en/, accessed 17 August 2018). 
  17. Ambient air pollution: A global assessment of exposure and burden of disease. Geneva, World Health Organization, 2016. 
  18. WHO Household Energy Database. Geneva: World Health Organization. (http://www.who.int/airpollution/data/household-energydatabase/en/, accessed 17 August 2018).

More information