WHO methods and data sources for life tables 1990-2015

Department of Information, Evidence and Research WHO, Geneva

May 2016

Global Health Estimates Technical Paper WHO/HIS/IER/GHE/2016.2 Acknowledgments

This Technical Report was written by Colin Mathers and Jessica Ho with inputs and assistance from Dan Hogan, Wahyu Retno Mahanani, Doris Ma Fat and Gretchen Stevens. WHO life tables were primarily prepared by Jessica Ho and Colin Mathers of the Mortality and Health Analysis Unit in the WHO Department of Information, Evidence and Research (in the Health Systems and Innovation Cluster of WHO, Geneva). Data and methods for the 2016 update of life tables were developed with advice and assistance from a WHO Lifetables Working Group, established by the WHO Reference Group on Global Health Statistics. We also drew on advice and inputs from Interagency Group on Child Mortality Estimation (UN IGME), the UN Population Division and UNAIDS. We would particularly like to note the assistance and inputs provided by Jeffrey Eaton, Patrick Gerland, Stephane Helleringer, Mary Mahy, Bruno Masquelier, Francois Pelletier, John Stover, John Wilmoth and Danzhen You.

Estimates and analysis are available at: http://www.who.int/gho/mortality_burden_disease/en/index.html http://www.who.int/gho

For further information about the estimates and methods, please contact [email protected]

In this series

1. WHO methods and data sources for life tables 1990-2011 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.1)

2. WHO-CHERG methods and data sources for child causes of death 2000-2011 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.2)

3. WHO methods and data sources for global causes of death 2000-2011 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.3)

4. WHO methods and data sources for global burden of disease estimates 2000-2011 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.4)

5. WHO methods for life expectancy and healthy life expectancy (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2014.5)

6. CHERG-WHO methods and data sources for child causes of death 2000-2012 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2014.6)

7. WHO methods and data sources for country-level causes of death 2000-2012 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2014.7)

8. MCEE-WHO methods and data sources for child causes of death 2000-2015 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2016.1) Contents

Acknowledgments...... Abbreviations ...... 1 Introduction ...... 1

2 WPP2015 life tables...... 1

3 General approach for preparation of annual life tables ...... 2

3.1 Interpolation of mx for countries in the WPP and VR categories ...... 4

3.2 Interpolation of mx for high HIV and “Other HIV” countries...... 6

4 Adjustments using death registration data ...... 20

4.1 Updated assessments of VR completeness ...... 40

4.2 Revision of imputed annual death rates ...... 42

5 Neonatal, infant and under five mortality ...... 133

6 WHO estimates of life expectancy and healthy life expectancy ...... 133

References ...... 155

Annex A: Data sources and methods for WHO Life Tables...... 177

Annex B: Data sources and methods for mortality shocks ...... 34

Annex C: Estimated completeness of death registration data ...... 41

Annex D: Estimated completeness of death registration data for most recent year ...... 48

Annex E: Comparison of 45q15 estimates, WPP2015 and GHE2016 ...... 49

ABBREVIATIONS

ARR Annual rate of reduction ART Anti-retroviral therapy CD Coale Demeny DHS USAID-supported Demographic and Health Surveys

EPP UNAIDS Epidemic Projection Package

GHE2015 WHO Global Health Estimates 2015 GHE2013 WHO Global Health Estimates 2013

ICD International Classification of Diseases IFHS Iraq Family Health Survey 2006-2007

IHME Institute for Health Metrics and Evaluation IMR Infant mortality rate

MICS UNICEF Multiple Indicator Cluster Surveys

mx age-specific death rates calculated from information on deaths among persons in the age group commencing at age x during a given time period and the total person-years for the population in the same age group during the same time period. For WHO and WPP2015 abridged life tables, all age groups from 5 onwards are 5-year age groups, m0 refers to infants aged 0 (first 12 months of life) and m1 refers to children aged 1 to 4 (ie. between exact ages 1 and 5).

NMR Neonatal mortality rate PCHIP Piecewise cubic Hermite interpolating polynomials PMTCT Prevention of Mother to Child Transmission of HIV PRIO Peace Research Institute Oslo

nqx probability of dying between exact ages x and x+n. U5MR Under-5 mortality rate UCDP Uppsala Conflict Data Program

UN-IGME Inter-agency Group for Child Mortality Estimation UNPD UN Population Division

VR Vital registration WHO World Health Organization

WHS 2002 to 2003 WHO World Health Survey program WPP2015 World Population Prospects 2015 1 Introduction

The World Health Organization (WHO) began producing annual life tables for all Member States in 1999. These life tables are a basic input to all WHO estimates of global, regional and country-level patterns and trends in all-cause and cause-specific mortality. After the publication of life tables for years to 2009 in the 2011 edition of World Health Statistics, WHO has shifted to a two year cycle for the updating of life tables for all Member States, and has moved towards alignment of this revision cycle with that of the World Population Prospects produced biennially by the UN Population Division.

Since 1998, WHO has been producing annual abridged life tables for Member States as part of its mandate to monitor and report on global progress in improving health. During the MDG era, WHO has been estimating time series of life tables from 1990 onwards. To support its reporting on progress towards the 2030 Agenda for Sustainable Development, WHO has released updated annual life tables for Member States for the period 1990-2015. These are available in the WHO Global Health Observatory (1) and in World Health Statistics 2016 (2). These updated life tables also provide the all-cause mortality estimates for the WHO Global Health Estimates 2015 (GHE2015) to be released in 2016.

In recent years, WHO has liaised more closely with the United Nations Population Division (UNPD) on life tables for countries, in order to maximize the consistency of UN and WHO life tables, and to minimize differences in the use and interpretation of available data on mortality levels. For almost all WHO Member States, this update draws on the World Population Prospects 2015 (WPP2015) life tables prepared by the UN Population Division (UNPD) (3), as well as on infant and under-5 mortality rates (U5MR) have been developed and agreed upon by the Inter-agency Group for Child Mortality Estimation (UN-IGME) which is made up of WHO, UNICEF, UNPD, World Bank and academic groups (4). death registration data reported to WHO by Member States (5), and UNAIDS/WHO estimates of HIV mortality for countries with high HIV prevalence (6).

The WHO Reference Group on Global Health Statistics convened a WHO Life Table Working Group in December 2014 to advise WHO on data, methods and revision strategies for WHO life tables. This group included academic experts as well as representatives of WHO, UNAIDS and UNPD and provided WHO with advice on the GHE2015 revisions through 2015 to date.

Consultations with Member States were carried out for estimates of neonatal, infant and child mortality in 2015, and for HIV mortality data, assumptions and estimates by UNAIDS in 2015 and 2016. Estimates for non-HIV mortality for ages 5 and over are closely aligned with the WPP2015 life tables, with adjustments for annual variations reported in death registration data and for mortality shocks (conflict and natural disasters). As a result, WHO did not carry out a separate consultation for the life table mortality rates, but will make these available to Member States for information, and for comments and new data to contribute to future revisions.

2. WPP2015 life tables

WPP2015 was released in mid-2015 (2) and includes abridged life tables for five-year periods 1950- 1955,……,2095-2100, for age groups 0, 1, 5, 10,….., 100+ by sex . Life tables are available for 183 of the 194 WHO Member States and for 3 non-Member territories with substantial populations (Occupied Palestinian Territory, Puerto Rico and China: Province of Taiwan). The WPP2015 excluded the following

World Health Organization Page 1

11 Member States (all with population <90,000 in 2015): Andorra, Cook Islands, Dominica, Marshall Islands, Monaco, Nauru, Niue, Palau, Saint Kitts and Nevis, San Marino, and Tuvalu. The most recent Global Health Estimates for causes of death (7) excluded 22 Member States with a population of less than 550,000 in 2015. For the current update, life tables have been prepared for the 183 WHO Member States included in WPP2015, as well as for the 3 largest non-Member territories. The latter will not be released, but included only for the calculation of regional and global life tables and all-cause mortality.

For a group of 21 “high HIV countries”, the WPP life tables were developed using Spectrum to model the HIV epidemic using Spectrum inputs and assumptions for HIV provided by UNAIDS in mid-2014. For these countries, UN Population Division has provided estimates of non-AIDS death rates in the form of model life tables indexed by e0 time series and specification of the model life table variant used for each country (mostly Coale Demeny North). This allows calculation of implied age-sex-specific HIV death rates for these countries.

3 General approach for preparation of annual life tables

For this update, the objective was to publish WHO annual period life tables for years 1990-2015. For internal use in other analyses, we also aimed to prepare annual life tables for 1985-1989.The starting point for the preparation of WHO annual life tables was to interpolate annual values for the age-specific

mortality rates mx from the WPP2015 5-year period average mx for each age-sex-country time series.

The WPP2015 uses the convention that a specified year (eg. 2015) refers to 1 July. Use of a hyphen (-) between years, for example, 1995-2000, signifies the full period involved, from 1 July of the first year to 1 July of the second year. WHO references to calendar years (eg 2015) refer to the period 1 January to 31 December and statistics either refer to averages or totals for the calendar year period. In practice, the annual average mortality rate for a calendar year is assumed to be essentially the same as the mortality rate at 1 July. Similarly, we assume that the 1 July population for year T is a proxy for the average population of calendar year T.

So if PT is the 1 July population for calendar year T, then the average population for the quinquennial period 2010-2015 is the person-years for the period 1 July 2010 to 1 July 2015 divided by 5:

Average population = (0.5*P2010 + P2012 + P2010 + P2013 +P2014 + 0.5*P2015)/5

For purposes of interpolation, we assume that the quinquennial average death rate falls in the centre of the quinquennial period eg. 1 January 2003 = 2002.5 This is usually an adequate approximation to the extent that trends are reasonably linear across the quinquennial interval.

For interpolation of mx values for annual periods 1985, 1986, …..2015 we used piecewise cubic Herite interpolating polynomials (usually referred to as PCHIP). This has the desirable property that he piecewise cubics join smoothly, so that both the interpolated function and its first derivative are continuous. In addition, the interpolant is shape-preserving in the sense that it cannot overshoot locally; sections in which period mx is increasing, decreasing or constant with time remain so after interpolation, and local extremes (maxima, maxima) also remain so (8). PCHIP interpolation was

World Health Organization Page 2

implemented using a procedure called pchipolate.ado available for Stata from the Statistical Software Components (SSC) Archive, often termed the "Boston College" archive (9).

When mx is not monotonically changing over time, and the mx for one period represents a local maximum or minimum, the interpolated annual mx will result in a period average mx that is lower than the local maximum input mx, or higher than the local minimum. This is illustrated in the following plot.

0.4 period average mx 0.35 3rd iteration 0.3 first iteration

0.25

0.2

0.15

0.1

0.05

0 1985 1990 1995 2000 2005 2010 2015 2020 2025

To ensure that the period average mx from the interpolated annual mx matches the inputs, the inputs were adjusted by the ratio of the original input to the new period average and the imputation repeated. Tests showed that three iterations were adequate to achieve reasonable convergence with average relative deviations in period mx below 0.001. We defined four groups of Member States for which data inputs and interpolation methods differed. The four groups are:

High HIV countries 21 countries for which WPP2015 used Spectrum to explicitly model HIV mortality. The UN Population Division has provided unpublished estimates of non-HIV mortality for these countries.

“Other” HIV countries An additional 22 countries with significant HIV epidemics for which WHO has in the past explicitly modelled HIV and non-HIV mortality trends in order to prepare life tables. These countries were not modelled using Spectrum for WPP2015.

VR countries 85 countries for which the WHO Mortality Database held mortality data from vital registration (VR) systems for 75% or more of years since 1990.

WPP countries 58 countries where interpolated mortality rates from WPP quinquennial life tables were used directly to construct annual life tables

A full list of countries in each category is provided in Annex Table A.

World Health Organization Page 3

3.1 Interpolation of mx for countries in the WPP and VR categories

Conflict and natural disasters (mortality shocks) may cause substantial increases in death rates for specific country-years. These may or may not be reflected in available death registration or survey/census data. WHO makes estimates of these deaths by country-year as part of its overall cause of death analyses and hence the all-cause mortality and life tables need to be consistent. Methods used for updating mortality shock estimates are summarized in Annex B.

The WPP 2015 includes the impact of large mortality shocks in some cases (eg. Rwanda 1994 genocide) but not for others (eg. Haiti 2010 earthquake). This may be obvious from 45q15 plots for large isolated mortality shocks, but much less clear for extended conflicts such as those in Afghanistan or Iraq. The assumed impact of mortality shocks included in WPP estimates may or may not be consistent with the WHO estimates of size of mortality shocks.

Annual estimates of conflict and natural disaster deaths by country, age, sex and year for the period 1985-2015 were prepared as described in Annex B. For countries in the VR and WPP groups, 45q15 plots were reviewed to identify country-periods for which WPP2015 had included an impact of mortality shocks. Separate plots for males and females were made for 45q15 calculated directly from the WPP 2015 estimates of mx and for “shock-free” mx from which the WHO estimates of shock mortality had been subtracted. The following plots show examples where the WPP 2015 estimates of mx included shock mortality (Bosnia and Herzegovina, left) and where they did not (Myanmar, right). For cases in the first category, non-shock mx were interpolated from the mx for neighbouring periods.

For six countries with extended conflicts, covering many of the 5-year periods in the range 1985-2015, it was assumed that the adult mortality data used to prepare WPP2015 life tables had included conflict mortality. These six countries were Afghanistan, Iraq, Sri Lanka, Somalia, Sudan and Yemen. For these countries, the WHO estimates of shock mortality for the 5-year periods were subtracted from the WPP2015 mortality rates and the 45q15 time series computed. The non-shock 45q15 were smoothed

World Health Organization Page 4 using Loess regression with bandwidth 0.8 and the smoothed 45q15 were used together with the UN model life table system (10,11) to compute non-shock mortality rates by age and sex.

Adjustments were also made for specific country-periods listed in Table 1, where WPP2015 had included some impact for a mortality shock. Table 2 lists country-periods with significant shock mortality according to WHO estimates, for which WPP2015 did not include a shock adjustment.

Table 1. Shocks identified as included in WPP 2015 mortality rates

WHO estimated

Country Period Type of shock Deaths/10,000 Deaths (‘000) Algeria 1990-2005 Conflict 8 32 Azerbaijan 1990-1995 Conflict 24 9 Bosnia and Herzegovina 1990-2000 Conflict 182 38 Croatia 1990-1995 Conflict 8 2 Georgia 1990-1995 Conflict 20 5 Indonesia 2000-2005 2004 Asian tsunami 14 166 Iran 1985-1990 Conflict and 1990 earthquake 13 40 2000-2005 Bam earthquake 2003 9 29 Lebanon 1985-1990 Conflict 52 7 Libya 2010-2015 Conflict 19 6 Mexico 1985-1990 1985 Mexico City earthquake 23 95 Micronesia 2000-2005 Tropical storm Chataan 2002 9 0.1 Nicaragua 1985-1990 Conflict 16 3 1995-2000 1998 Hurricane Mitch 14 3 Palestine 1985-2015 Conflict 12 12 Pakistan 2005-2010 Conflict + 2005 Kashmir earthquake 13 106 2010-2015 5 47 Panama 1985-1990 Disaster 15 2 Peru 1990-1995 Conflict 7 8 Rwanda Samoa 2005-2010 Disaster 17 0.2 Syria 2010-2015 Conflict 274 269 Tajikistan 1990-1995 Conflict 44 12 Turkey 1995-2000 1999 Izmit earthquake 14 43 Venezuela 1995-2000 1999 floods and mudslide 26 30

World Health Organization Page 5

Table 2. Shocks identified as NOT included in WPP 2015 mortality rates

WHO estimated

Country Period Type of shock Deaths/10,000 Deaths ('000) Armenia 1985-1990 1988 earthquake 105 37 Bangladesh 1990-1995 1991 Bangladesh cyclone 25 142 Columbia 1985-1990 1985 volcanic eruption 16 23 Comoros 1995-2000 Conflict 6 147 Croatia 1995-2000 Conflict 4 1 El Salvador 1985-1990 Conflict 135 34 Georgia 2005-2010 Conflict 6 1 Honduras 1995-2000 1998 Hurricane Mitch 49 15 Kuwait 1990-1995 Conflict 236 22 Myanmar 2005-2010 Cyclone Nargis 2008 56 141 Nepal 2000-2005 Conflict 16 20

WPP2015 estimates of adult mortality for Lebanon were largely based on trends in U5MR and CD West model life tables. Data on adult mortality from reported household deaths in some recent censuses and surveys in the regions suggested that adult mortality is often higher than implied by CD west for a given U5MR, and adult mortality rates for Lebanon were adjusted upwards based on the region-specific relationship between 45q15, u5mr and income per capita. Additionally, Lebanon, Turkey and Jordan are three countries with very substantial de facto resident populations of Syrian refugees from 2013 onwards. For years 2013-2015, mortality rates for these three countries were adjusted using a population weighted average of WPP2015 estimates of the country mx and Syrian mx, using UNHCR estimates of refugee populations (12).

After preparation of a complete set of “shock-free” period mx for the VR and WPP countries, annual mx were interpolated and WHO annual estimates of shock mortality added to obtain total mortality rates by 5-year age group, sex and year from 1985-2015.

3.2 Interpolation of mx for high HIV and “Other HIV” countries

For countries with substantial proportion of younger adult deaths (15-60 years) due to HIV, the all-cause mortality envelopes, trends and age patterns must be consistent with the HIV mortality estimates, otherwise the “non-HIV envelopes” will have strange and implausible age and/or time trends. This will then affect most other cause of death estimates.

For development of previous WHO life tables, 43 countries were classified as “high HIV” and explicit efforts made to ensure consistency of all-cause and HIV mortality estimates. For the WPP 2015, UN Population Division used Spectrum (13)with input assumptions consistent with those of UNAIDS in mid- 2014 to model all-cause mortality for 21 countries.

World Health Organization Page 6

Following discussions at the WHO life table working group meeting in New York, October 2015, Avenir Consulting prepared updated Spectrum models for 1985-2015 that took into account the WPP 2015 revisions to demographic data and all-cause mortality, as well as 2015 UNAIDS files with a range of 2016 updates to the Spectrum/AIM software including new patterns of adult mortality on ART and age at ART initiation among pediatric patients and the re-fitting of all the EPP curves.

Among the most important Spectum/AIM updates were:

1. Improvements to the EPP model that fits smooth prevalence trends to surveillance and survey data. The handling of entrants and exists from the population 15-49 was improved to reduce the differences between the EPP model, which is a single age/sex group, and AIM which divides the population by sex and single age.

2. The mother-to-child transmission rates were updated by reviewing new studies from the last three years. This resulting in changes to the probability of transmission from mother-to-child for some PMTCT regimens.

3. New patterns of mortality for adults on ART were developed by the IeDEA Consortium using new data from 2011 to 2014. These updated patterns show somewhat higher mortality than the patterns used last year mainly as a result of the inclusion of more countries in the IeDEA data set. 4. The pediatric model now has a pattern describing the age distribution of children newly starting ART. The pattern was derived from data from IeDEA treatment sites. Previously new ART patients were distributed by age according to need. The new patterns are regional and show shifts in the distribution by age over time as infant diagnosis has expanded.

For preparation of the WHO life tables, it was also necessary to address issues relating to large mortality shocks in some countries, and the need to exclude these before interpolating from 5-year period to annual mortality rates. The following table summarizes WHO analyses of mortality shocks likely included in the mx estimates for the 43 high HIV and other HIV countries.

Table 3. Shocks identified as included in WPP 2015 mortality rates for HIV countries

WHO estimated

Country Period Type of shock Deaths/10,000 Deaths ('000) Angola 1985-1990 Conflict 36 19 High HIV 1990-1995 Conflict 65 39 High HIV Burundi 1995-2000 Conflict 36 12 High HIV 2000-2005 Conflict 28 10 High HIV Chad 1985-1990 Conflict 16 4 Other HIV 1990-1995 Conflict 14 5 Other HIV 1995-2000 Conflict 6 2 Other HIV 2000-2005 Conflict 6 3 Other HIV 2005-2010 Conflict 12 6 Other HIV Congo 1995-2000 Conflict 191 28 High HIV DR Congo 1995-2000 Conflict 33 75 Other HIV Rwanda 1990-1995 Conflict 1567 512 High HIV

World Health Organization Page 7

Table 4. Shocks identified as NOT included in WPP 2015 mortality rates for HIV countries

WHO estimated

Country Period Type of shock Deaths/10,000 Deaths ('000) Angola 1985-1990 Conflict 36 19 High HIV Cameroon 1985-1990 1986 Lake Nyos Gas Disaster 0.3 2 High HIV Djibouti 1990-1995 Conflict 17 1 Other HIV Central African Republic 2010-2015 Conflict 25 6 High HIV Eritrea 1995-2000 Conflict 224 37 Other HIV 2000-2005 Conflict 66 13 Other HIV Ethiopia 1985-1990 Conflict 23 50 High HIV 1990-1995 Conflict 15 39 High HIV Haiti 2005-2010 2010 Earthquake 14 188 Other HIV

Table 5. UN Model life table systems (UNMLT) used for non-HIV mortality estimates in HIV countries

High HIV countries UNMLT Other HIV countries UNMLT Angola CD North Burkina Faso CD North

Burundi CD North Côte d'Ivoire CD North Botswana CD West Ghana CD North Central African Republic CD North Guinea CD North

Cameroon CD North Haiti CD North Congo CD North Liberia CD North

Ethiopia CD North Mali CD North Gabon CD North Nigeria CD North Equatorial Guinea CD North Chad CD North

Kenya CD North Togo CD North Lesotho CD West Thailand UN Far_East_A sian Mozambique CD North Malawi CD South Namibia CD West Rwanda CD North Swaziland CD West United Republic of Tanzania CD North Uganda CD North South Africa UN Far_East_Asian Zambia CD North Zimbabwe CD North

World Health Organization Page 8

For high HIV countries, provisional non-HIV mx were calculated from the model life table assumptions and e0 series provided by UN Population Division. We added estimates of HIV death rates based on the revised Spectrum models to the non-HIV mx to recomputed total mortality mx. This led to consequential changes in trends and/or levels of all-cause 45q15 for a number of countries. To reduce these differences and to smooth trends for non-HIV mortality, revised model life tables for non-HIV mortality were prepared for 8 countries: Central African Republic, Ethiopia, Gabon, Kenya, Lesotho, Malawi, Mozambique, and Rwanda. The model life table assumptions for the high HIV countries are shown in Table 5. Some adjustments to individual period mx were also required for 8 countries to take into account mortality shocks. In the case of South Africa, all-cause death registration data adjusted for completeness was also used to assess levels of all-cause mortality, resulting in HIV mortality estimates somewhat lower than UNAIDS and WPP2015 estimates.

For “other HIV countries”, we subtracted the revised Spectrum modelled HIV mortality rates from the WPP2015 all-cause mortality rates and examined the consistency and plausibility of the resulting non- HIV mortality time trends, age trends and sex ratios. Provisional WHO mx were calculated by smoothing the implied non-HIV mortality trends and adding back the UNAIDS HIV mortality estimates. For problems with isolated periods in 8 countries, the mx for the period were revised by interpolation. For the following 11 countries, adjusted 45q15 time series were used to revise the model life tables for non-HIV mortality: Burkina Faso (females only), Chad, Côte d'Ivoire, Ghana, Guinea, Haiti, Liberia, Mali, Nigeria, Togo and Thailand (males only). The model life table assumptions for the “Other HIV” countries are shown in Table 5.

Figure 3. Comparison of HIV mortality rates for 2010 for Scatterplots of resulting 45q15 (non-HIV and GHE2015 versus GHE2013 total) identified implausibly low mortality levels 600 for Tanzania in most recent years. The most

HIV_high recent publicly available empirical mortality 500 "HIV_med" Lesotho pattern available for Tanzania (based on

2015)

HE 400 reported household deaths in the 2012 census) Swaziland (G was also consistent with a higher 45q15 than the 000

, 300 100

WPP2015 estimate. This derived from an

er p

s Mozambique estimated rapid decline of child mortality in h 200

eat d Zambia recent years, used to predict adult non-HIV IV H mortality from the model life table for WPP2015. 100 The most recent publicly available empirical

0 mortality pattern available for Tanzania (based 0 100 200 300 400 500 600 700 on reported household deaths in the 2012 HIV deaths per 100,000 (GHE2013) census) was also consistent with a higher 45q15 than the WPP2015 estimate. Trends in adult 45q15 for Tanzania were revised to reduce the acceleration in rate of decline, drawing also on IHME analyses of non-HIV mortality for Tanzania (14).

World Health Organization Page 9

4 Adjustments using death registration data

The WPP2015 life tables draw extensively on available death registration data to assess age-specific mortality rates mx. For 21 countries with high quality and complete death registration data, they make use of life tables prepared for the Human Mortality Database (15, 16), which corrects for age misstatement at older age groups.

WHO holds time series of death registration data for around 100 countries (5). These potentially provide alternate data for preparing annual life tables, or additional data that would assist in imputation from period life tables. For 85 countries with at least 75% of the years in range 1990-2015 available, we evaluated the completeness of the all-cause deaths data and used completeness-adjusted death rates to inform the imputation of annual death rates for life tables.

4.1 Updated assessments of VR completeness

Until now, WHO has used population data reported by Member States for the population covered by death registration as the denominator for calculation of all-cause mortality rates, and has then computed total numbers of deaths by applying these rates to WPP population estimates for the de-facto resident population. This can result in completeness estimates varying from 100% for high income countries, and to estimated total deaths lower than registered deaths. Additionally, country-reported population data are not available for a substantial proportion of country-years in some regions.

For this revision of completeness estimates, we have switched to use of WPP2015 population estimates as denominators. Implied completeness of death registration data has then been assessed against WPP 2015 by comparing reported registered deaths against the total deaths implied by the WPP life tables for 5-year periods.

There are five countries where registered deaths reported to WHO do not include a province or territory not under government control. These are:

Cyprus: all data refer to government controlled areas Georgia: excluding Abkhazia and South Osetia Moldova: excluding Transnistria and Bender Russia: 1993-2003 data exclude Chechnya

Serbia: excluding Kosovo-Metohija province

Singapore does not report deaths for non-citizen residents, who represent approximately 30% of the de- facto population. A number of European countries similarly exclude non-citizen residents from data reported to WHO those these typically amount to at most a few percent of the de-facto resident population. For these two groups of countries, the new method potentially results in lower completeness against UN estimates of resident population.

Overall completeness levels for ages 15+ were estimated by sex and compared with completeness estimates from IHME derived using an ensemble of death distribution methods (17,18) and with previous WHO completeness estimates based on WHO application of the Generalized Brass Growth- Balance and Bennett-Horiuchi methods for the 1990s and early 2000s. For the 85 countries with VR time

World Health Organization Page 10

series, adult completeness was assessed by comparing total registered deaths for persons aged 15 years and over in each five year period with 5-year period deaths for ages 15+ calculated from the WPP2015 life table mx together with WPP2015 population estimates for the 5-year periods. For 5-year periods covering only 4 years of death registration data, total registered deaths for the missing year was interpolated or extrapolated. For 5-year periods covering 2 or 3 years of death registration data, completeness was assessed against the total deaths calculated from the corresponding annual lifetables interpolated from WPP2015. In most cases, where a beginning or end period contained only 1 year of death registration data, this was excluded from the analysis.

Adjustments to estimated completeness levels were made for some countries as follows:

Guyana: Prior to 1995, IHME estimates of completeness were used

Israel: inclusion of East Jerusalem from 1980 onwards would result in apparent completeness of 1.06- 1.07 against WPP2015. Completeness was truncated at 1.0

Malta: completeness assessed against WPP2015 exceeded 1.0 for all except most recent years. Completeness was truncated at 1.0.

Mauritius: completeness assessed against WPP2015 exceeded 1.0 for all except most recent years. The deaths reported to WHO include all of Mauritius except for the island of Rodrigues. Completeness for Mauritius was thus revised to 0.974.

Russia: Deaths in Chechnya are missing from the data reported to WHO for years 1993-2003 (corresponding to approximately 1% incompleteness). Assessed completeness was slightly less than 0.99 for most of this period, and was revised to 1.0 for males in 1990-1995 period, when it slightly exceeded 1.0.

Suriname: Prior to 1995-2000, previous WHO estimates of completeness were used

For a number of other countries mainly in Eastern Europe and Latin America and the Caribbean, completeness estimates fluctuated both above and below 1.0. Apparent completeness above 100% may reflect issues with the numerator for registered deaths, which can vary in some country-years in term of (a) "year of occurrence vs year of registration", (b) "provisional vs. final", (c), de-jure vs. de-facto or (d) inclusion of nationals only (as in Japan and several EU countries). It may also reflect mismatches with denominators resulting from issues around the estimation of the "de-facto" population and the inclusion of migrants and refugees in the countries (irrespective of their legal status). For country- periods where estimated completeness exceeded 1.05, it was capped at 1.05.

After these adjustments, annual completeness estimates were smoothly interpolated from the 5-year period completeness estimates. Rising and falling projections for latest partial 5-year period were adjusted to avoid out-of-sample trends. In a number of cases where completeness rates for the last 5- year period rose above 100%, completeness was capped at 100%. Identification of out of sample trends was also informed by examination of IHME estimates of completeness trends (18). Annex C compares the resulting final completeness time series for countries with previous estimates by WHO and IHME. Annex Table D lists estimated completeness for the most recent year of death registration data for each Member State meeting inclusion criteria.

World Health Organization Page 11

4.2 Revision of imputed annual death rates

The WPP2015 life tables based on death registration data included adjustments for age mis-statement and under-reporting at older ages based on analyses carried out by the Human Mortality Database (15,16), analyses for consistency with population age structures and use of the Kannisto-Thatcher method for assessing oldest age mortality rates (19). For this reason, we did not simply apply the completeness estimated for ages 15+ to all VR deaths for those ages, but carried out a second assessment of completeness against the WPP2015 life tables, for each sex separately for age groups 5- 9,10-14, 15-69, 70-79, 80-84, 85+.

Smoothly varying annual completeness estimates for years 1985-2015 were imputed using PCHIP interpolation (8,9). The age-sex specific completeness estimates were assumed constant outside the lower and upper period mid-points for the available VR years for each country. Age-specific completeness estimates for age groups 70-74 and 75-79 were then imputed to ensure a smooth transition between earlier and older age groups. Finally, the imputed age-sex specific completeness estimates were adjusted to match the sex-specific completeness for ages 15+. Completeness for age groups 5-9 and 10-14 was capped at 100%.

For 7 countries with less than 1 million population in 2000-2015, VR death rates were smoothed using a 3-year moving average. Annual mx for ages 5+ were calculated for each VR country-year as:

mx = VR deaths /(WPP population estimate)/(annual age-sex-specific completeness estimate)

While most VR countries had data up to 2012 or 2013, only a few had reported data to WHO for 2014, and none for 2015 at the time of analysis (early 2016). VR-based mx estimates for age groups from 5-9 onwards were projected forward using Poisson regression to estimate the trend in latest 10 years of VR data. Importance weights declining by a factor of 0.85 for each earlier year from last were used to give greater weight to more recent trend. Estimated annual rates of change (ARR) for 5-year age groups were smoothed across age groups using a 3-age group moving average. ARRs not statistically significant were capped within the range of statistically significant ARRs across age groups. The mx values were projected forward to 2015 using an ARR that changed smoothly from the VR-based ARR estimate to the WPP2015- based ARR over a six year period. For most countries, this means that the 2015 estimated mx are quite strongly influenced by the trends in the recent VR data, where for the handful of countries where the latest year was earlier than 2009, the VR trend converged to the WPP trend over the project period.

Annual mx for all age groups from 5-9 onwards were replaced by the VR-based estimates and projections for VR countries for which completeness analyses were carried out with the exception of the following five countries: Singapore, Cyprus, Kuwait, Malaysia and Sri Lanka These five countries had partial VR coverage and/or major fluctuations in implied completeness estimates (see Annex C).

World Health Organization Page 12

5 Neonatal, Infant and Under-five mortality Mortality rates for infants and age group 1-4 years for the WHO life tables were derived from the UN- IGME estimates of infant mortality rates (IMR) and under 5 mortality rates (U5MR) by sex, for Member States for years 1990-2015 (20). The United Nations Inter-agency Group for Child Mortality Estimation (UN IGME), which includes UNICEF, WHO, the World Bank and United Nations Population Division, was established in 2004 to advance the work on monitoring progress towards the achievement of Millennium Development Goals regarding child mortality.

UN-IGME annually assesses and adjusts all available surveys, censuses and vital registration data, to estimate country-specific trends in neonatal (NMR), infant (IMR) and under 5 (U5MR) mortality rates per 1,000 live births. All data sources and estimates are documented on the website www.childmortality.org. For countries with complete recording of child deaths in death registration systems, these are used as the source of data for the estimation of trends in neonatal, infant and child mortality. For countries with incomplete death registration, all other available census and survey data sources, which meet quality criteria, are used.

Due to fewer data available by sex than data for both sexes, UN IGME uses available data by sex to estimates time trend in the sex ratio (male/female) of U5MR. Leontine Alkema and Fengqing Chao of the National University of Singapore have developed new Bayesian methods for the UN IGME estimation of sex ratios, with a focus on the estimation and identification of countries with outlying levels or trends (21).

6 WHO estimates of life expectancy and healthy life expectancy

6.1 Life expectancy

Final estimates of age-sex-specific mortality rates for years 1990-2015 were used to compute abridged life tables for 183 WHO Member States with population of 90,000 or greater in 2015. Life expectancies at birth were reported in World Health Statistics 2016: and full life tables are available in the WHO Global Health Observatory (www.who.int/gho). Annex E presents country plots showing the resulting WHO annual estimates of 45q15 by sex for all-cause mortality and for non-HIV mortality excluding disasters and conflict deaths. Five year period estimates of 45q15 from the WPP2015 life tables are shown for comparison.

WHO applies standard methods to the analysis of Member State data to ensure comparability of estimates across countries. This will inevitably result in differences for some Member States with official estimates for quantities such as life expectancy, where a variety of different projection methods and other methods are used. These WHO estimates of mortality and life expectancies should not be regarded as the nationally endorsed statistics of Member States, which may have been derived using alternative methodologies and assumptions.

INSERT SUMMARY OF DATA AVAILABILITY HERE

There remain substantial data gaps and deficiencies in data on levels of child and adult mortality, particularly in those regions with the highest mortality levels. Quantifiable uncertainty ranges for adult mortality are more complex to derive, and there is considerable research underway to develop

World Health Organization Page 13

improved methods for measuring adult mortality in surveys, and in assessing the systematic biases in such data. Table 6 summarizes the availability of data on levels of all-cause mortality for WHO Member States and the methods used to assess mortality and life expectancy.

Table 6 Data availability for all-cause mortality

Number of WHO Percentage of global a b Available recent data (since 2005) Member States deaths 2015 Methods c Complete death-registration data 59 28 Observed death rates Incomplete death-registration data 38 25 Adjusted death rates Other population-representative data on age- 18 (3) 25 Estimated death rates and d specific mortality model life table systems Data on child (under 5 years) and 30 (18) 12 Estimated death rates and d adult (15–59 years) mortality only model life table systems d Data on child mortality only 37 (22) 10 Model life table systems No recent data 1 <1 Projected from data for years before 2005 a Only includes 183 Member States with population above 90 000 in 2015. b Percentage of global deaths that occur in the countries included in each category – not the percentage registered or included in datasets. c Completeness of 90% or greater for de facto resident population; as assessed by WHO and the United Nations, Population Division, 2016. d Numbers in parenthesis show the number of high HIV prevalence countries for which multistate epidemiological modelling for HIV mortality was also carried out.

A qualititative guide to the uncertainty in adult mortality and life expectancy estimates is provided by the listing of methods and data input types in Annex Table A. The most reliable estimates are those based on death registration data assessed as complete, followed by those based on incomplete or sample death registration data with adjustments for levels of completeness. For countries without useable death registration data, uncertainties are substantially higher, and two categories can be distinguished (a) those countries where there is independent evidence on adult mortality rates from surveys or censuses and (b) those where estimates of adult mortality levels are derived from model life tables with estimated infant and child mortality rates as inputs. Those countries with significant levels of mortality due to conflict and natural disasters (say, greater than 1 death per 10,000 population per annum) will usually have additional uncertainty associated with the difficulties in estimating conflict and disaster death rates.

6.2 Healthy life expectancy

WHO has previously published estimates of healthy life expectancy (HLE or HALE) for years 2000 and 2012, drawing on the previous WHO life table series and estimates of years lost to disability (YLD) for disease and injury causes from the Global Burden of Disease 2010 (GBD2010) study (22-24).

The same methods have been used to prepare estimates of healthy life expectancy for WHO Member States for the year 2015 (2), using the updated WHO life tables and projections of YLD for year 2015 based on YLD estimates for 2010 and 2013 from the Global Burden of Disease 2013 (GBD2013) study(25), with similar adjustments to disability weights and prevalences for certain causes as previously (22).

World Health Organization Page 14

References

(1) The Global Health Observatory (GHO) is WHO’s portal providing access to data and analyses for monitoring the global health situation. See: http://www.who.int/gho/en/ (http://www.who.int/gho/mortality_burden_disease/life_tables/en/index.html)

(2) World Health Statistics 2016. Geneva: World Health Organization; 2016. (http://apps.who.int/iris/bitstream/10665/206498/1/9789241565264_eng.pdf)

(3) World Population Prospects: The 2015 Revision. DVD Edition. New York (NY): UnitedNations, Department of Economic and Social Affairs, Population Division; 2015 (http://esa.un.org/unpd/wpp/)

(4) Levels & Trends in Child Mortality. Report 2015. Estimates Developed by the UN Interagency Group for Child Mortality Estimation. New York (NY), Geneva and Washington (DC): United Nations Children’s Fund, World Health Organization, World Bank and United Nations; 2015 (http://www.unicef.org/publications/files/Child_Mortality_Report_2015_Web_9_Sept_15.pdf).

(5) World Health Organization. Mortality Database. Available at: http://www.who.int/healthinfo/mortality_data/en/index.html

(6) UNAIDS (2015). HIV estimates with uncertainty bounds 1990-2014. Available at 2015http://www.unaids.org/en/resources/documents/2015/HIV_estimates_with_uncertainty_bounds_1 990-2014 (accessed 19 May 2016)

(7) Global Health Estimates 2013: deaths by cause, age and sex; estimates for 2000–2012. Geneva: World Health Organization; 2014 (http://www.who.int/healthinfo/global_burden_disease/en/).

(8) For more information, see http://blogs.mathworks.com/cleve/2012/07/16/splines-and- pchips/#ee54c20b-0ecc-4ac9-8986-8a0774a1763f

(9) Boston College Department of Economics. PCHIPOLATE: Stata module for piecewise cubic Hermite interpolation. Statistical Software Components no. S457561. Available at http://fmwww.bc.edu/repec/bocode/p/pchipolate.ado

(10) Coale AJ, P Demeny and B Vaughan. 1983. Regional Model Life Tables and Stable Populations. New York: Academic Press.

(11) UN Population Division. 2010. World Population Prospects 2012: Extended Model Life Tables. New York: United Nations, Department of Economic and Social Affairs. http://esa.un.org/wpp/Model-Life- Tables/data/MLT_UN2010-130_1y.xls

(12) UNHCR (2015). Syria Regional Refugee Response: Inter-agency Information Sharing Portal. Available at http://data.unhcr.org/syrianrefugees/regional.php

(13) Stover J, Andreev K, Slaymaker E, et al. Updates to the Spectrum model to estimate key HIV indicators for adults and children. AIDS (London, England). 2014;28(4):S427-S434. doi:10.1097/QAD.0000000000000483.

(14) GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and national age–sex specific all- cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2014; 385: 117–71.

(15) Human Mortality Database. University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available at www.mortality.org or www.humanmortality.de

(16) Wilmoth JR, Andreev KF, Jdaov DA, Glei DA. Methods Protocol for the Human Mortality Database. Version 5. University of California, Berkeley and Max Planck Institute for Demographic Research; 2007.

World Health Organization Page 15

(17) Murray CJL, Rajaratnam JK, Marcus J, Laakso T, Lopez AD (2010) What Can We Conclude from Death Registration? Improved Methods for Evaluating Completeness. PLoS Med 7(4): e1000262. doi:10.1371/journal.pmed.1000262

(18) Phillips DE, Rafael Lozano R, Mohsen Naghavi M, Charles Atkinson C, Diego Gonzalez-Medina D, Lene Mikkelsen L, Christopher JL Murray CJL, Alan D Lopez AD. A composite metric for assessing data on mortality and causes of death: the vital statistics performance index. Population Health Metrics 2014, 12:14. DOI: 10.1186/1478-7954-12-14

(19) UNICEF, WHO, The World Bank and UN Population Division. Levels and Trends of Child Mortality - Report 2015, Estimates developed by the UN Inter-agency Group for Child Mortality Estimation. UNICEF, New York, 2015

(20) UNICEF, WHO, The World Bank and UN Population Division. Levels and Trends of Child Mortality - Report 2015, Estimates developed by the UN Inter-agency Group for Child Mortality Estimation. UNICEF, New York, 2015

(21) Alkema L, Chao F, Sawyer CC (2013). Sex Differences in U5MR: Estimation and identification of countries with outlying levels or trends. Paper presented at the XXVII IUSSP International Population Conference, Busan, Republic of Korea.

(22) World Health Organization (2014). WHO methods for life expectancy and healthy life expectancy (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2014.5). Available at www.who.int/evidence/bod

(23) Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M et al (2012a). Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet.380:2163–2196.

(24) Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M et al (2012b). Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010 [Supplementary appendix]. (http://download.thelancet.com/mmcs/journals/lancet/PIIS0140673612617292/mmc1.pdf?id=a02f57d18 11fcb77:-1b44796c:142333b8265:-259e1383841102443, accessed 7 November 2013)

(25) Global Burden of Disease Study 2013 Collaborators: Vos T et al. Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet. 2015 June 8. doi: 10.1016/S0140-6736(15)60692-4.

World Health Organization Page 16

Annex Table A: Data sources and methods for WHO Life Tables

WPP2015 methods for estimation of age-sex-specific mortality GHE2015 WHO MBD VR Member State ratesa Life table method method years available Afghanistan Estimated using the West model of the Coale-Demeny Model Life Tables CD West relational wpp - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adjusted mortality estimates of adult mortality were derived from: (a) recent household deaths data from the 1979 census; (b) implied relationship between child mortality and adult mortality based on the UN South Asian and West model of the Coale-Demeny Model Life Tables; and (c) levels of adult mortality based on sample registration data from neighbouring countries for recent years. Estimates of adult mortality derived from (i) recent household deaths data from the 2010 Afghanistan Mortality Survey (AMS), (ii) parental orphanhood from the 2010 AMS (excluding the Southern region), and (iii) siblings deaths from the 2010 AMS (excluding the Southern region) adjusted for age misreporting and recall biases were also considered.

Angola Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Albania Based on life tables for 1987-2013 derived from registered deaths by age Death registration data vr 1980, 1984-2009 and sex and observed trends in infant and child mortality.

United Arab Based on life tables derived from official estimates of registered deaths and Death registration data wpp 2003, 2005-2010 Emirates enumerated census population by age and sex from 1988 through 2010, adjusted for infant and child mortality. Mortality rates for older ages were adjusted.

Argentina Based on registered deaths from 1950 through 2013, and the underlying Death registration data vr 1980-2013 population from censuses, and revised projections by the National Statistics Office (INDEC). The number of deaths was adjusted using the growth- balance method.

Armenia Based on: (a) a life table derived from reported deaths by age and sex in Death registration data vr 1981-2012 2011 and the 2011 census population, adjusted for underreporting of infant and child deaths, and (b) official estimates of life expectancy available from 2006 through 2013.

Antigua and Based on official estimates of life expectancy from 2000 to 2010. Death registration data wpp 1983-2009, 2012- Barbuda 2013

Australia Based on official estimates of life expectancy available through 2009. The Death registration data vr 1980-2012 age pattern of mortality is based on life tables through 2009 from the Human Mortality Database.

Austria Based on official estimates of life expectancy at birth through 2013. Death registration data vr 1980-2014

Azerbaijan Based on deaths registered through 2012 classified by age and sex and the Death registration data vr 1981-2011 underlying population by age and sex. Death rates were adjusted for underregistration.

Burundi Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV model life tables, and taking into account the number of deaths due to civil mortality strife. The demographic impact of AIDS has been factored into the mortality estimates.

Belgium Based on official estimates of life expectancy available through 2012. Death registration data vr 1980-2012

Benin Estimated using the North model of the Coale-Demeny Model Life Tables CD North relational Other HIV - and implied relationships between life expectancy at birth and estimates of model for non-HIV infant and child mortality, and between life expectancy at birth and mortality estimates of adult mortality (45q15). The adjusted estimates of 45q15 were derived from: (a) parental orphanhood from the 1981/83 multiround survey and 2002 census, and (b) siblings deaths from the 1996, 2002 and 2006

World Health Organization Page 17

DHS.

Burkina Faso Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Data from mortality West African rural demographic surveillance sites and urban vital registration were also considered. Adjusted estimates of adult mortality were derived from: (a) recent household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic- extinct generation methods) from the 1960/61 survey, 1976, 1985, 1996 and 2006 censuses, the 1991 National Demographic Survey, and 2008 Global Fund survey; (b) parental orphanhood from the 1993, 2003 and 2010/11 DHS, 2006 MICS3 and 2006 census; (c) siblings deaths from the 1998/99, 2003 and 2010/11 DHS; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1976-1985, 1985-1996 and 1996-2006.

Bangladesh Based on life tables derived from age and sex-specific mortality rates from: Death registration data wpp 1980-1982, 1984- (a) the Sample Vital Registration System from 1981 up to 2010 adjusted for 1986 infant and child mortality, (b) the 1974 Retrospective Survey of Fertility and Mortality, and (c) the 1962/65 Population Growth Estimation Experiment. Estimates are consistent with those from the 2001 and 2010 Bangladesh Maternal Mortality Surveys (based on sibling histories and household deaths in the preceding 36 months), and data gathered from Matlab Health and Demographic Surveillance System up to 2012. For the period 1970- 1975, mortality was adjusted to take into account the excess mortality associated with the 1971 civil war and independence from Pakistan, and the 1974 flood and famine.

Bulgaria Based on official life tables through 2013. Death registration data vr 1980-2012

Bahrain Based on life tables derived from official estimates of registered deaths and Death registration data wpp 1980-2013 enumerated census population by age and sex from 1980 to 2012, adjusted and CD South model life for infant and child mortality. Mortality rates for older ages were adjusted. tables For the period 1950-1980, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South model of the Coale-Demeny Model Life Tables in 1950-1955, and converges over time toward the estimated 1980-1985 life table.

Bahamas Derived from child and adult mortality estimates through 2013 by assuming Death registration data Other HIV 1980-2012 that the age pattern of mortality conforms to the West model of the Coale- and CD West relational Demeny Model Life Tables. model, adjusted for UNAIDS estimates of HIV mortality

Bosnia and Based on official estimates of life expectancy for 1988/89 and WHO Death registration data vr 1985-1991, 1998- Herzegovina estimates for years 2000 to 2012. The estimates of war-related deaths in 2011 the period 1992-1995 were also considered.

Belarus Based on official life tables available through 2013. Death registration data vr 1981-2009, 2011- 2012

Belize Estimated using the West model of the Coale-Demeny Model Life Tables and Death registration data Other HIV 1980-2013 two parameters: (a) estimates of child mortality; and (b) adjusted estimates and CD West relational of adult mortality from registered deaths and underlying population model, adjusted for through 2009. From 1950 to 1995, estimated using adjusted registered UNAIDS estimates of HIV deaths by age and sex and underlying population by age and sex. mortality

Bolivia Based on life tables derived from: (a) deaths by age and sex, adjusted using Survey, census and vr 2000-2003 (Plurinational the growth-balance method, and underlying population from the 1992, death registration data State of) 2001 and 2012 censuses; (b) data on maternal orphanhood from the 1988 National Population and Housing Survey (ENPV); (c) official estimates of life expectancy for 2010 and 2011; and (d) estimates of infant and child mortality from the 2000 MICS and the 1989, 1994, 1998, 2003 and 2008 DHS.

Brazil Based on life tables derived from: (a) registered deaths by age and sex from Death registration data vr 1980-2013 1979 through 2012, and the underlying census population by age and sex, and (b) estimates of infant and child mortality. The number of deaths was

World Health Organization Page 18

adjusted using the growth-balance method.

Barbados Derived from estimates of child mortality and adult mortality from vital Death registration data vr 1980-2012 registration data through 2007 by assuming that the age pattern of and CD West model life mortality conforms to the West model of the Coale-Demeny Model Life tables Tables.

Brunei Derived from child and adult mortality estimates through 2011 by assuming CD West relational vr 1982-2013 Darussalam that the age pattern of mortality conforms to the West model of the Coale- model for non-HIV Demeny Model Life Tables. Life tables are estimated using the Flexible two- mortality dimensional model life table and Lee-Carter method.

Bhutan Based on a life table derived from adjusted deaths in the past 12 months by Survey, census data and wpp - age and sex, and the population by age and sex from the 2005 census, CD West Model life adjusted for infant and child mortality. For 1950-2000, life tables were tables derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the estimated 2000-2005 life table. Life tables based on adjusted annual deaths from the 1994 National Health Survey were also considered.

Botswana Derived from estimates of infant and child mortality by assuming that the CD West model life High HIV 1995 age pattern of mortality conforms to the West model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Central African Estimated using the North model of the Coale-Demeny Model Life Tables CD North relational High HIV 1988 Republic and implied relationships between life expectancy at birth and estimates of model for non-HIV infant, child, and adult (45q15) mortality. The adjusted estimates of adult mortality mortality were derived from (a) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for the period of 1988-2000; (b) household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic-extinct generation methods) from the 1988 census; (c) parental orphanhood data from the 1994/95 DHS and the 1988 census; and (d) siblings deaths from the 1994/95 DHS. The demographic impact of AIDS has been factored into the mortality estimates.

Canada Based on official estimates of life expectancy available through 2011. The Death registration data vr 1980-2011 age pattern of mortality is based on life tables through 2011 from the Human Mortality Database.

Switzerland Based on official life tables from through 2012. Death registration data vr 1980-2012

Chile Based on life tables derived from registered deaths, and population by age Death registration data vr 1980-2013 and sex from 1950 to 2013 adjusted for infant and child mortality. The number of deaths was adjusted using the growth-balance method,

China Based on life tables from: (a) the 1981, 1990, 2000 and 2010 censuses Survey, census and wpp 1995-2000 (adjusted for underestimation of child mortality and overestimation of old- sample death age mortality); (b) surveys on causes of death in 1973/75 and 2004/05; (c) registration data Disease Surveillance Points (DSP) system from 1991 to 2013; and (d) 1987, 1995 and 2005 population survey (1 per cent), and the annul survey on population change (1 per thousand).

Côte d'Ivoire Estimated using the North model of the Coale-Demeny Model Life Tables CD North relational Other HIV 1998 and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates are derived from (a) recent household deaths data from the 1978/79 follow-up survey, the 1998 census and the 2005 EIS, (b) parental orphanhood from the 1978/79 follow-up survey, the 1988 and 1998 censuses, the 1994 DHS, the 2000 MICS2 and the 2006 MICS3 surveys, (c) siblings deaths from the 1994 DHS and the 2005 EIS; (d) implied relationship between child mortality and adult mortality based on the South model of the Coale-Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the North model of the Coale-Demeny Model Life Tables by the 1970s.

Cameroon Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV 1987 age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

World Health Organization Page 19

Democratic Derived from (a) estimates of infant and child mortality by assuming that CD North model life Other HIV - Republic of the the age pattern of mortality conforms to the North model of the Coale- tables Congo Demeny Model Life Tables, and (b) data on survival of siblings from the 2007 and 2013/14 DHS.

Congo Derived from estimates of infant and child mortality by assuming that the CD West model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Colombia Based on life tables derived from registered deaths, and population by age Death registration data vr 1982-2012 and sex from 1950 to 2013, adjusted for infant and child mortality. The number of deaths was adjusted using the growth-balance method.

Comoros Derived from estimates of infant and child mortality, and the West model of CD West model life wpp - the Coale-Demeny Model Life Tables. tables

Cabo Verde Derived from estimates of child mortality, by assuming that the age pattern CD West model life wpp - of mortality conforms to the West model of the Coale-Demeny Model Life tables Tables. Official estimates of life expectancy at birth by sex for 1990 and 2000 were also considered.

Costa Rica Based on life tables derived from registered deaths, adjusted using the Death registration data vr 1980-2013 growth-balance method, and population by age and sex from 1950 to 2013 adjusted for infant and child mortality.

Cuba Based on: (a) deaths registered through 2012 classified by age and sex and Death registration data vr 1980-2013 the underlying population by age and sex, and (b) estimates of infant and child mortality. The number of deaths was adjusted using the growth- balance method.

Cyprus Based on: (a) official life tables; (b) deaths registered through 2013 Death registration data wpp 1980-2012 classified by age and sex and on the underlying population by age and sex; and (c) estimates from other areas were also considered.

Czech Republic Based on official estimates of life expectancy available through 2013. The Death registration data vr 1982-2013 age pattern of mortality is based on official life tables for 1950-2013.

Germany Based on official estimates of life expectancy available through 2013. Death registration data vr 1980-2013

Djibouti Derived from estimates of infant and child mortality by assuming that the CD North model life Other HIV 1991 age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables.

Denmark Based on official life tables available through 2012. Death registration data vr 1980-2012

Dominican Based on: (a) registered deaths by age and sex through 2011, and Survey, census and vr 1980-1992, 1994- Republic underlying mid-year population; (b) estimates of infant and child mortality death registration data 2012 from 2000, 2006, and 2014 (preliminary) Multiple Indicator Cluster Survey (MICS); (c) estimates of infant and child mortality from the 1986, 1991, 1996, 2002, 2007 and 2013 DHS. The number of deaths was adjusted using the growth-balance method.

Algeria Based on official estimates of life expectancy derived from the number of Death registration data wpp 1980-1982, 1985- deaths registered through 2013. Estimates were adjusted for under- 1986, 1998, 2000 reporting of deaths and deaths of non-nationals. From 1950 to 2000, the age patterns of mortality are derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the mortality patterns resulting from the blending from the South model of the Coale-Demeny Model Life Tables to the West model of the Coale-Demeny Model Life Tables from 1950 to 2000.

Ecuador Based on: (a) registered deaths by age and sex from 1954 to 2011, with Death registration data vr 1980-2013 underlying mid-year population; (b) estimates from the 1989, 1994, 1999 and 2004 ENDEMAIN, and the 1987 ENDESA; and (c) estimates from the 1950, 1962, 1974, 1982, 1990, 2001 and 2010 censuses. The number of deaths was adjusted using the growth-balance method.

Egypt Based on official estimates of life expectancy available through 2013. The Death registration data vr 1980-1981, 1983- age pattern of mortality is based on official life tables for various years from 2013 1960 to 2012 adjusted for infant and child mortality, and adult mortality.

Eritrea Derived from estimates of infant and child mortality by assuming that the CD North model life Other HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables.

World Health Organization Page 20

Spain Based on: (a) official estimates of life expectancy available through 2013; Death registration data vr 1980-2013 (b) registered deaths by age and sex through 2012 and underlying population by age and sex; (c) estimates from the Human Mortality Database; and (d) estimates from Eurostat were also considered.

Estonia Based on official life tables available through 2013. Death registration data vr 1981-2012

Ethiopia Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV 1984 age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Finland Based on official life tables available through 2013. Death registration data vr 1980-2013

Fiji Based on: (a) official 1976, 1986, 1996, 2001 and 2007 estimates, and (b) Death registration data wpp 1980-1987, 1992- deaths by age and sex registered from 1950 through 2007. 2009, 2011-2012 France Based on official life tables through 2012. vr 1980-2012

Micronesia Based on estimates of infant and child mortality by assuming that the age CD West model life wpp - (Federated pattern of mortality conforms to the West model of the Coale-Demeny tables States of) Model Life Tables.

Gabon Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

United Kingdom Based on: (a) official life tables for 2010-2012, and (b) life tables derived Death registration data vr 1980-2013 from official estimates of registered deaths and population from 1950 to 2011.

Georgia Based on official estimates of life expectancy available through 2012, Death registration data vr 1981-2014 adjusted for underregistration.

Ghana Estimated using the South model of the Coale-Demeny Model Life Tables CD North relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from: (a) recent female household deaths data from the 2007 Ghana Maternal Health Survey; (b) parental orphanhood from the 1988, 1993, 1998, 2003 and 2008 DHS as well as 2006 MICS3 survey; (c) siblings deaths from the 2007 Ghana Maternal Health Survey; and (d) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables.

Guinea Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15) derived from mortality (a) recent household deaths data from the 1954-1955 Demographic Survey, and the 1983 and 1996 censuses; (b) parental orphanhood from the 1999 and 2005 DHS ; (c) siblings deaths from the 1999, 2005 and 2012 DHS ; (d) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the South model of the Coale- Demeny Model Life Tables by the 1990s. Data from West African rural demographic surveillance sites and urban vital registration were also considered, including from the 1957 Urban Survey (Konkoure).

Gambia Estimated using the South model of the Coale-Demeny Model Life Tables and CD South relational Other HIV - three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality was derived from the relationship to child mortality implied by the North model of the Coale-Demeny Model Life Tables. Adult mortality estimates derived from recent household deaths data from the 1973 census, and from parental orphanhood from the 1973, 1983 and 2003 censuses, 2001 Baseline Survey in Lower, Central and Upper River Divisions, 2000 and 2005/06 MICS surveys and rural demographic surveillance sites were also considered. The results of the 2013 DHS were considered.

World Health Organization Page 21

Guinea-Bissau Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from (a) parental orphanhood from the 2000 and 2006 MICS surveys, and (b) implied relationship between child mortality and adult mortality based on the North model of the Coale- Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1990s. Data from West African rural demographic surveillance sites (e.g., Bandim) and urban vital registration were also considered.

Equatorial Estimated using the North model of the Coale-Demeny Model Life Tables CD North relational High HIV - Guinea and implied relationships between life expectancy at birth and estimates of model for non-HIV infant and child mortality and between life expectancy at birth and mortality estimates of adult mortality (45q15). The adjusted estimates of 45q15 were derived from (a) parental orphanhood data from the 2000 MICS; and (b) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for the period of 1983-1994. The demographic impact of AIDS has been factored into the mortality estimates.

Greece Based on official life tables available through 2012. Death registration data vr 1980-2012

Grenada Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 1985-2013 age pattern of mortality conforms to the West model of the Coale-Demeny tables Model Life Tables.

Guatemala Based on: (a) registered deaths by age and sex and the underlying mid-year Death registration data vr 1980-1981, 1983- population by age and sex through 2013; (b) age-sex-specific death rates 2013 from the 1995, 1998/99, 2002 and 2008/09 Encuestas Nacionales de Salud Materno Infantil (ENSMI); (c) age-sex-specific death rates from the 1987 and 1989 Encuestas Nacionales Socio-demográficas (ENSD); and (d) death rates by age and sex from the 1950, 1964, 1973, 1981, 1994 and 2002 censuses. The number of deaths was adjusted using the growth-balance method.

Guyana Derived from child and adult mortality estimates through 2010 by assuming CD West relational vr 1984, 1988-2011 that the age pattern of mortality conforms to the West model of the Coale- model for non-HIV Demeny Model Life Tables. Adult mortality estimates based on the Global mortality Burden of Disease Study 2010 were considered.

Honduras Based on: (a) registered deaths by age and sex and the underlying mid-year Survey, census and wpp 1980-1983, 1987- population by age and sex from 1950 through 1983 and from 2000 through death registration data 1990, 2008-2013 2011; (b) ages-sex-specific death rates from the 2005/06 and 2011/12 ENDESA (DHS); (c) ages-sex-specific death rates from the 1991/92, 1996 and 2001 ENESF; (d) ages-sex-specific death rates from the 1987 EFHS, the 1984 MCH/PF, the 1971/72 and 1983 EDENH, the 1981 National Contraceptive Prevalence Survey (EPAH); and (e) estimates from the 1974, 1988 and 2001 censuses. The number of deaths was adjusted using the growth-balance method.

Croatia Based on deaths registered through 2013 by age and sex and the underlying Death registration data vr 1982-2013 population by age and sex.

Haiti Based on: (a) estimates from the 1987, 1994/95, 2000 and 2005/06 DHS; (b) Survey, census and Other HIV 1980, 1997, 1999, estimates from the 1982 and 2003 censuses, (c) registered deaths by age and death registration data 2001-2004 sex adjusted for incompleteness using the growth-balance method and the 1971 census population by age and sex; and (d) estimates from the 1977 Enquête Haitienne sur la Fécondité (EHF).

Hungary Based on official estimates of life expectancy available through 2013. The Death registration data vr 1980-2013 age pattern of mortality is based on official life tables for through 2013.

Indonesia Derived from estimates of infant, child, adult and old-age mortality. Adult CD North relational wpp - and old age mortality estimates are based on: (a) the 2002/03, 2007 and model for non-HIV 2012 DHS, (b) the 1990, 2000 and 2010 censuses, and (c) the 2007/08 mortality Indonesia Family Life Survey (IFLS), and (d) the SUSENAS surveys (National Socio-economic Surveys).

India Based on life tables derived from age and sex-specific mortality rates from Death registration data wpp 1988-2008 the Sample Registration System from 1968-1969 up to 2013 adjusted for infant and child mortality, and for adult death underregistration by using the growth-balance and synthetic-extinct generation methods.

Ireland Based on official estimates of life expectancy through 2009. Death rates Death registration data vr 1980-2012 estimated from 2010 to 2012 were also considered.

World Health Organization Page 22

Iran (Islamic Based on life tables derived from age and sex-specific mortality rates from Death registration data wpp 1983-1984, 1986, Republic of) (a) registered 2000-2012 annual deaths adjusted for infant and child 1991, 1995-1999, mortality, and for adult death underregistration using the growth-balance 2001, 2005-2008 and synthetic-extinct generation methods; (b) the 1956-1966 intercensal survival, 1973/76 Population Growth Survey, 1976, 1986 and 1991 censuses, and 2000 Demographic and Health Survey; and (c) estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables. For 1980-1988, excess mortality due to the war was factored in the overall mortality levels based on the PRIO Battle Deaths Dataset version 3.0, released in October 2009.

Iraq Estimated using the West model of the Coale-Demeny Model Life Tables CD West model life wpp 1987-1989, 2008 and three parameters: (1-2) direct and indirect estimates of infant and child tables mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality estimates are derived from: (a) recent household deaths data from the 1973/74 Demographic Sample Survey and Sample Registration, and 1999 Child and Maternal Mortality Survey (female only); (b) parental orphanhood from the 1997 census, 2004 Iraq Living Conditions Survey and 2006 MICS3; (c) siblings deaths from the 1990 Iraq Immunization, Diarrhoeal Disease, Maternal and Childhood Mortality Survey (female only), and the 2006/07 Iraq Family Health Survey; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1957-1965, 1965-1977, 1977-1987 and 1987-1997; (e) implied relationship between child mortality and adult mortality based on the West model of the Coale-Demeny Model Life Tables. For 1980-1988, excess mortality due to the war was factored in the overall mortality levels based on the PRIO Battle Deaths Dataset version 3.0, released in October 2009. For 2000-2005 and 2005-2010, excess mortality due to the war was factored in the overall mortality levels; there is a high level of uncertainty in the current estimates. The estimated numbers of war related deaths, as provided by the Iraqi Ministry of Health, have also been taken into account.

Iceland Based on official life tables available through 2013. Death registration data vr 1980-2012

Israel Based on life tables derived from official estimates of registered deaths and Death registration data vr 1980-2014 enumerated census population by age and sex from 1948 to 2013. Mortality rates for older ages were adjusted.

Italy Based on: (a) life tables for 2010, 2011 and 2012 from the National Death registration data vr 1980-2012 Statistical Office (Istat) and Eurostat; (b) life tables through 2005-2009 from the Human Mortality Database.

Jamaica Based on: (a) registered deaths by age and sex through 2005, adjusted for Census and death Other HIV 1980-1991, 1996- underreporting of infant and child deaths; (b) official estimates for 1991, registration data 2011 2002, 2003 and 2006; and (c) estimates from the 2001 and 2011 censuses.

Jordan Estimated using the West model of the Coale-Demeny Model Life Tables CD West model life wpp 2003-2004, 2008- and three parameters: (1-2) direct and indirect estimates of infant and child tables 2011 mortality, and (3) estimates of adult mortality (45q15) implied by the relationship between child mortality and adult mortality based on the South model of the Coale-Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the West model of the Coale-Demeny Model Life Tables by the 1980s. Life tables based on the 1961 and 1979 censuses, 1972 National Fertility Survey and 1976 WFS, indirect estimates of adult mortality based on widowhood data from the 1961 and 1979 censuses and 1976 WFS, as well as parental orphanhood from the 1976 WFS and 1981 Demographic Survey were also taken into account.

Japan Based on life tables derived from official estimates through 2012. Death registration data vr 1980-2013

Kazakhstan Based on official estimates of life expectancy available through 2008 Death registration data vr 1981-2010, 2012 adjusted for underreporting of infant and child mortality. The age pattern of mortality is derived from a life table based on 2010-2013 data.

World Health Organization Page 23

Kenya Estimated using the North model of the Coale-Demeny Model Life Tables CD North relational High HIV - and implied relationships between life expectancy at birth and estimates of model for non-HIV infant and child mortality and between life expectancy at birth and mortality estimates of adult mortality (45q15). The adjusted estimates of 45q15 were derived from: (a) household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic-extinct generation methods) from the 1969, 1979, 1989, 1999 and 2009 censuses; (b) parental orphanhood from the 1983, 1989, 1993, 1998, 2003, and 2008/09 Kenya DHS, the 1977 World Fertility Survey, and all censuses aforementioned; (c) siblings deaths from the 1989, 1993, 1998 2003, and 2008/09 Kenya DHS; and (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for the period of 1969-2009. The demographic impact of AIDS has been factored into the mortality estimates.

Kyrgyzstan Based on official estimates of life expectancy available through 2013 Death registration data vr 1981-2013 adjusted for underreporting of infant and child mortality.

Cambodia Based on life tables derived from age and sex-specific mortality rates from: Survey, census data and wpp - (a) recent household deaths data from the 2004 Inter-Censal Population CD West Model life Survey and 2008 census; (b) siblings deaths from the 2000, 2005 and 2010 tables DHS. Also, 1950-1955 life tables derived from estimates of child mortality were used, by assuming that the age pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables, as well as recent household deaths data from the 1959 rural survey, and the 1962-1998 population reconstructions.

Kiribati Based on: (a) estimates in the 2005 and 2010 censuses; (b) estimates from CD West model life wpp 1991-2001 deaths by age and sex from 1995 to 2001; (c) child mortality estimates by tables assuming that the age pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables; and (d) estimates from the Secretariat of the Pacific Community were also considered.

Republic of Based on official estimates of life expectancy through 2012. Death registration data vr 1980-2013 Korea

Kuwait Based on life tables derived from official estimates of registered deaths and Death registration data wpp 1980-1989, 1991- enumerated census population by age and sex from 1964 to 2010, adjusted 2013 for infant and child mortality. Mortality rates for older ages were adjusted. For 1950-1965, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the estimated 1965-1970 life table.

Lao People's Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 1995 Democratic age pattern of mortality conforms to the West model of the Coale-Demeny tables Republic Model Life Tables.

Lebanon Derived from estimates of infant and child mortality by assuming that the CD West model life wpp - age pattern of mortality conforms to the West model of the Coale-Demeny tables Model Life Tables. For the period 2010-2015, life expectancy at birth was adjusted to account for different mortality patterns of large Syrian refugee population.

Liberia Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15) derived from mortality (a) recent household deaths data from the 1969-1970 and 1970-1971 Population Growth Surveys ; (b) parental orphanhood from the 2007 DHS ; (c) siblings deaths from the 2007 DHS ; (d) implied relationship between child mortality and adult mortality based on the West model of the Coale- Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1990s. Data from West African rural demographic surveillance sites and urban vital registration were also considered.

Libya Derived from estimates of infant and child mortality by assuming that the CD West model life wpp - age pattern of mortality conforms the East model of the Coale-Demeny tables Model Life Tables and converges over time toward the West model of the Coale-Demeny Model Life Tables from 1950 to 2010.

World Health Organization Page 24

Saint Lucia Based on: (a) official estimates of life expectancy available through 2005; Death registration data vr 1980-2006, 2008- (b) registered deaths by age and sex through 2005 and underlying 2012 population by age and sex; and (c) estimates from the 1991 and 2001 censuses.

Sri Lanka Based on: life tables derived from official estimates of registered deaths and Death registration data wpp 1980-2007 population by age and sex from 1950 to 2010, adjusted for infant and child mortality, and for adult death underregistration for males before 1980 by using tabulations of paternal orphanhood (before marriage) by age of respondent from the 1987 Sri Lanka DHS.

Lesotho Derived from estimates of infant and child mortality by assuming that the CD West model life High HIV - age pattern of mortality conforms to the West model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Lithuania Based on official estimates of life expectancy available through 2011. The Death registration data vr 1980-2013 age pattern of mortality is based on life tables through 2011 from the Human Mortality Database.

Luxembourg Based on official estimates of life expectancy available through 2012. The Death registration data vr 1980-2013 age pattern of mortality is based on official life tables through 2012.

Latvia Based on official life tables available through 2012. Death registration data vr 1980-2012

Morocco Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 1991-1998, 2008- age pattern of mortality initially conforms to the mortality patterns tables 2012 resulting from the blending from the East model of the Coale-Demeny Model Life Tables to the West model of the Coale-Demeny Model Life Tables from 1950 to 2015.

Republic of Based on official estimates of life expectancy available through 2012 Death registration data vr 1981-2013 Moldova adjusted for underreporting of infant and child mortality. The age pattern of mortality is derived from a life table based on data for 2012.

Madagascar Based on: (a) estimates from the 1966 Demographic Survey and the 1973 CD North relational wpp - and 1993 censuses; (b) estimates derived from registered age-sex-specific model for non-HIV deaths and underlying age-sex-specific population; (c) estimates derived mortality from implied relationships between child mortality and adult mortality from the 1992, 1997, 2003/04 and 2008/09 DHS based on the North model of the Coale-Demeny Model Life Tables; and (d) 1966 estimates from OECD were also considered.

Maldives Based on life tables derived from official estimates of registered deaths and Death registration data vr 1984-2011 enumerated census population by age and sex from 1975 to 2012, adjusted for infant and child mortality and for adult death underregistration for males in 1980-1985 using the growth-balance and synthetic-extinct generation methods. For 1950-1975, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South-Asian model of the United Nations Model Life Tables in 1950-1955 and converges over time toward the estimated 1975-1980 life table.

Mexico Based on: (a) registered deaths by age and sex through 2013 and underlying Death registration data vr 1980-2013 population by age and sex, (b) estimates from the 1992, 2006 and 2009 Encuesta Nacional de la Dinámica Demográfica (ENADID), (c) estimates from the1978 and 1979 ENPUMA, and the 1976 WFS, (d) estimates from the 1970, 1990, 2000 and 2010 censuses. The number of deaths was adjusted using the growth-balance method.

The former Based on official estimates of life expectancy available through 2012. The Death registration data vr 1982-2010 Yugoslav age pattern of mortality is based on an official life table for 2010-2012. Republic of Macedonia

World Health Organization Page 25

Mali Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV 1987 and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15) derived from mortality (a) recent household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic-extinct generation methods) from the 1957-1958 Demographic Survey (Central Delta) and 1960-61 Demographic Survey, the 1976, 1987, 1998 and 2009 censuses; (b) parental orphanhood from the 1995-1996, 2001 and 2006 DHS ; (c) siblings deaths from the 1995-1996, 2001, 2006 DHS and 2012- 2013 DHS; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1976-1987, 1987- 1998, and 1998-2009 ; (e) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables in 1950-1955, and assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1980s.

Malta Based on official life tables through 2012. Death registration data vr 1980-2014

Myanmar Derived from estimates of infant and child mortality by assuming that the CD West model life wpp - age pattern of mortality conforms to the West model of the Coale-Demeny tables Model Life Tables. Official estimates of life expectancy at birth by sex from the 1991 Myanmar Population Change and Fertility Survey and the recent household deaths data from the 2014 census were also considered.

Montenegro Based on official estimates of life expectancy available through 2010. The Death registration data vr 1985-2010 age pattern of mortality is based on life tables for 1990, 2000 and 2006.

Mongolia Based on official estimates of life expectancy available through 2013 Death registration data vr 1991-2010 adjusted for underreporting of infant and child mortality.

Mozambique Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Mauritania Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational wpp 1988 and implied relationships between life expectancy at birth and estimates of model for non-HIV infant and child mortality, and estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from: (a) parental orphanhood from the 2007 MICS3, 2000/01 DHS, 1964/65 Demographic Survey and 1981/82 Fertility Survey of Mauritania (WFS), (b) siblings deaths from the 2000/01 DHS, (c) household deaths data from the 1957 Fouta Toro survey, 1977 and 1988 censuses, and (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1965-1977, 1977- 1988 and 1988-2000.

Mauritius Based on official life tables through 2013. Death registration data vr 1980-2014

Malawi Derived from estimates of infant and child mortality by assuming that the CD South model life High HIV 1987, 1998 age pattern of mortality conforms to the South model of the Coale-Demeny tables for non-HIV Model Life Tables. Estimates from the 1987, 1998 and 2008 censuses and mortality official estimates from the National Statistical Office of Malawi were also considered. The demographic impact of AIDS has been factored into the mortality estimates.

Malaysia Based on: (a) official estimates of life expectancy available through 2008, Death registration data wpp 1990-2009 and (b) 2011 and 2012 official estimates of deaths by age and sex and the underlying population by age and sex.

Namibia Derived from estimates of infant and child mortality by assuming that the CD West model life High HIV - age pattern of mortality conforms to the West model of the Coale-Demeny tables for non-HIV Model Life Tables. Official estimates from Statistics Namibia were also mortality considered. The demographic impact of AIDS has been factored into the mortality estimates.

World Health Organization Page 26

Niger Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational wpp - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from (a) recent household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic-extinct generation methods) from the 1959/60 Demographic Survey, 1977, 1988 and 2001 censuses; (b) parental orphanhood from the 1988 and 2001 censuses, 1992 and 1998 DHS, 2006 DHS-MISC3; (c) siblings deaths from the 1992 DHS, 2006 DHS-MICS3 and 2012 DHS-MICS4; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1977-1988, and 1988-2001; (e) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables in 1950-1955, and assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1990s. Data from West African rural demographic surveillance sites and urban vital registration were also considered.

Nigeria Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from: (a) recent household deaths data from the 1965-1966 Nigerian rural demographic inquiry, the 2008 and 2013 DHS, and the 2010/11 GHS; (b) parental orphanhood from the 1986, 1999, 2003, 2008 and 2013 DHS, the 2007 MICS3 and 2010/11 GHS; (c) siblings deaths from the 2008 DHS; (d) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables. Data from West African rural demographic surveillance sites including for Malumfashi in 1962-1966 and 1974-1977 and urban vital registration were also considered.

Nicaragua Based on: (a) registered births and infant and child deaths from 1968 Death registration data vr 1987-1994, 1996- through 2011; (b) estimates from the 1998, 2001, 2006/07, and the 2013 2011/12 (preliminary) ENDESA (DHS); (c) estimates from the 1992/93 Family Health Survey, the 1993 and 2001 National Household Survey on Living Standards Measurement (LSMS); the 1985/86 National Socio-Demographic Survey, the 1978 National Retrospective Demographic Survey; and (d) estimates from the 1953, 1963, 1971, 1995 and 2005 censuses. The number of deaths was adjusted using the growth-balance method.

Netherlands Based on official estimates of life expectancy derived from registered Death registration data vr 1980-2013 deaths through 2013. The age pattern of mortality is based on official life tables for 1950 to 2013.

Norway Based on official life tables available through 2013. Death registration data vr 1980-2013

Nepal Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 1981, 1991 age pattern of mortality conforms to the West model of the Coale-Demeny tables Model Life Tables.

New Zealand Based on official estimates of life expectancy available through 2009. The Death registration data vr 1980-2011 age pattern of mortality is based on life tables through 2008 from the Human Mortality Database.

Oman Based on life tables derived from official estimates of registered deaths for Death registration data wpp 2009-2010 2009-2011 and 2010 enumerated census population by age and sex, adjusted for infant and child mortality. For 1950-2007, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the estimated 2009-2011 life table.

Pakistan Based on life tables derived from age and sex-specific mortality rates from: Survey, census and wpp 1984-1993 (a) the 1962-1965 Population Growth Estimation Experiment, 1968-1971 sample death Population Growth Survey I, 1976-1979 Population Growth Survey II; (b) the registration data 1984-2007 annual Pakistan Demographic Surveys adjusted for infant and child mortality, and for adult death underregistration for males in 1950- 1970 using the growth-balance and synthetic-extinct generation methods, as well as cross-validation with other countries experiencing similar mortality levels; and (c) estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South-Asian model of the United Nations Model Life Tables. Estimates are consistent with those based on parental survival and widowhood data from the 1984 PDS. Mortality rates for older ages were adjusted.

World Health Organization Page 27

Panama Based on: (a) registered deaths by age and sex and underlying population Death registration data vr 1980-2013 by age and sex from 1952 through 2013; (b) estimates from the 1950, 1970, 1980, 1990, 2000, and 2010 censuses; and (c) official life tables for 1960, 1970, 1979, 1989 and 1999. The number of deaths was adjusted using the growth-balance method.

Peru Based on: (a) registered deaths by age and sex through 2012 and the Survey, census and vr 1980-1992, 1994- underlying population by age and sex; (b) official estimates in 1961, 1965, death registration data 2013 1980, 1990, 1995, 2000, and 2005; (c) estimates of infant and child mortality from 2004-2014 continuous Encuestas Demográficas y de Salud Familiar (ENDES/DHS), and the 1986, 1991/92, 1996 and 2000 ENDES; (d) estimates of infant and child mortality from the 1977/78 World Fertility Survey, the 1974/76 National Demographic Survey; and (e) estimates from the 1961, 1972, 1981, 1993, and 2007 censuses. The number of deaths was adjusted using the growth-balance method.

Philippines Based on: (a) child mortality estimates from the 1998, 2003, 2006 and Death registration data vr 1980-2005, 2007- 2013 DHS, and 2006 Family Planning Survey, (b) estimates od infant and 2009 child mortality, (c) official estimates from a life table of 2006, and (d) the West model of the Coale-Demeny Model Life Tables and the Lee-Carter method.

Papua New Based on: (a) infant and child mortality estimates, (b) parental survivorship UN Far Eastern relational wpp 1987-1998 Guinea (orphanhood) data by age from the 2000 census, (c) child mortality data model for non-HIV from the 1996 and 2006 PNG DHS, by assuming that the age pattern of mortality mortality conforms to the Far Eastern model of the United Nations Model Life Tables.

Poland Based on official estimates of life expectancy available through 2013. The Death registration data vr 1980-2013 age pattern of mortality is based on official life tables through 2013.

Puerto Rico Based on: (a) registered deaths by age and sex through 2014 and underlying Death registration data vr 1980-2013 population by age and sex, and (b) official estimates of life expectancy available through 2010.

Democratic Based on the number of deaths in household during the 12-month period Census data wpp 1993 People's preceding the 1993 and 2008 censuses classified by age and sex. Republic of Korea

Portugal Based on: (a) official estimates of life expectancy available through 2012; Death registration data vr 1980-2013 (b) registered deaths by age and sex through 2011 and underlying population by age and sex; and (c) estimates from the Human Mortality Database and Eurostat were also considered.

Paraguay Based on: (a) registered deaths by age and sex through 2006 and underlying Survey, census and wpp 1980-1987, 1990, population by age and sex; (b) estimates from the 2004 and 2008 ENDSSR death registration data 1992, 1994-2013 and the1995/96 ENDSR; (c) estimates from the 2003 WHS, the 1998 National Maternal and Child Health Survey, the 1990 DHS, the 1987 RHS, the 1979 WFS, and the 1977 National Demographic Survey; and (d) estimates from the 1950, 1962, 1972, 1992, and 2002 censuses, and preliminary results from the 2012 census. The number of deaths was adjusted using the growth-balance method.

Occupied Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 2011 Palestinian age pattern of mortality conforms to the West model of the Coale-Demeny tables Territory Model Life Tables.

Qatar Based on life tables derived from official estimates of registered deaths and Death registration data wpp 1981-1983, 1985- enumerated census population by age and sex from 1981 to 2011, adjusted 2012 for infant and child mortality. Mortality rates for older ages were adjusted. For 1950-1980, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the estimated 1980-1985 life table.

Romania Based on official life tables through 2012. Death registration data vr 1980-2012

Russian Based on official estimates of life expectancy available through 2012. The Death registration data vr 1980-2011 Federation age pattern of mortality is based on life tables through 2012 from the Human Mortality Database. Both estimates incorporate an adjustment to infant mortality.

World Health Organization Page 28

Rwanda Based on the estimated level of infant mortality and taking into account the CD North model life High HIV - unusual numbers of deaths caused by the 1993-1994 civil war. The tables for non-HIV demographic impact of AIDS has been factored into the mortality estimates. mortality

Saudi Arabia Based on official estimates of life expectancy at birth for 2010-2013. For Death registration data wpp 2009, 2012 1995-2010, based on life-tables, calculated from adjusted deaths in the past 12 months by age and sex, and the population by age and sex from the 1999 Demographic Survey, 2004 census and 2007 Demographic Survey adjusted for infant and child mortality, and old-age mortality. For 1950- 1995, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the South model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the West model of the Coale-Demeny Model Life Tables and the estimated 1999-2007 life tables. Life tables based on annual deaths from the 2000 Demographic Survey, and on 2005 and 2009 registered deaths were also considered.

Sudan Derived from estimates of infant and child mortality by assuming that the CD North model life wpp - age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables.

Senegal Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational wpp - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from (a) recent household deaths data (unadjusted and adjusted for underregistration using the growth-balance and synthetic-extinct generation methods) from the 1978/79 Multiround Survey, 1988,2002 and 2013 censuses; (b) parental orphanhood from these sources and the 1986, 1992/93, 2005 DHS and 2010/11 DHS-MICS, 1988 census, and 2000 MICS; (c) siblings deaths from the 1992/93, and 2005 DHS and 2010/11 DHS-MICS; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1976-1988 and 1988-2002; (e) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables in 1950-1955, assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1990s; and (f) central deaths rate by age from the 2013 census.

Singapore Based on: (a) official estimates of life tables through 2010, and (b) 2011- Death registration data wpp 1980-2014 2013 official estimates of deaths and population by age and sex.

Solomon Islands Based on: (a) data on children ever born and surviving from the 1986 and CD West relational wpp - 1999 censuses; (b) official estimates based on census analysis and WHO- model for non-HIV GBD estimates for 2006; and (c) 1980-1984 life table based on indirect mortality methods assuming that the pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables..

Sierra Leone Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15). Adult mortality mortality estimates were derived from: (a) recent household deaths data from the 1992 Demographic and social monitoring survey and the 2004 census; (b) parental orphanhood from the 1973 pilot census. 1974, 1985 and 2004 censuses, 2000 MICS2, 2005 MICS3, 2007 CWIQ and 2008 DHS surveys; (c) female sibling deaths from the 2005 MICS3, and sibling deaths from the 2008 and 2013 DHS; (d) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for periods 1963- 1974, 1974-1985, 1985-2004; (e) implied relationship between child mortality and adult mortality based on the South model of the Coale- Demeny Model Life Tables for males, and the North model for females for the 1950-1970 period. Data from West African rural demographic surveillance sites (including from the 1973/75 Ad-hoc survey in Greater Freetown, the Western area and Makeni in the Northern Province) and urban vital registration were also considered.

El Salvador Based on: (a) registered deaths from 1975 through 2008, and underlying Death registration data vr 1980-2012 population by age and sex; (b) estimates from the 1950, 1963, 1971, 1992 and 2007 censuses; (c) estimates from the 1973 to 2008 Encuesta Nacional de Salud Familiar (FESAL), the 1992 EHS, and the 1985 DHS. The number of deaths was adjusted using the growth-balance method.

Somalia Derived from estimates of infant and child mortality by assuming that the CD North model life wpp - age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables. Estimates from GBD-WHO were also considered. Additional deaths due to the famine of 1992 and the war have been

World Health Organization Page 29

factored into the mortality estimates.

Serbia Based on official estimates of life expectancy available through 2011. The age Death registration data vr 1985-2013 pattern of mortality is based on official life tables for 1997, and for 2005 to 2012.

South Sudan Derived from estimates of infant and child mortality by assuming that the CD North model life Other HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables.

Sao Tome and Based on: (a) official estimates and life table derived from the 2001 census; CD North relational wpp 1984-1985, 1987 Principe and (b) death rates calculated from registered deaths by age and sex model for non-HIV through 1979 and underlying population by age and sex. Estimates mortality from WHO-GBD and estimates derived from child and adult mortality using North Model of the Coale-Demeny Model Life Table were also considered.

Suriname Based on: (a) registered deaths by age and sex through 2013 and underlying Death registration data vr 1980-2012 population by age and sex, and (b) official estimates for 1963, 1980, 2004 and 2006.

Slovakia Based on official life tables through 2013. Death registration data vr 1982-2014

Slovenia Based on official estimates of life expectancy available through 2012. The Death registration data vr 1982-2010 age pattern of mortality is based on blended life tables (from the East model of the Coale-Demeny Model Life Tables assumed to convert over time toward the empirical data in 1980) between 1950 and 1980, and official life tables from 1980 to 2012.

Sweden Based on official life tables available through 2013. Death registration data vr 1980-2014

Swaziland Derived from estimates of infant and child mortality by assuming that the CD West model life High HIV - age pattern of mortality conforms to the West model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Seychelles Based on: (a) official estimates available through 2014, and (b) registered Death registration data wpp 1980-2014 deaths by age and sex through 2014 and underlying population by age and sex.

Syrian Arab For 2005-2010, based on a life-table calculated from 2005-2007 registered Death registration data wpp 1983-1984, 1998, Republic deaths by age and sex, and post-censal population estimates by age and sex 2000-2001, 2004, derived from the 2004 census and 2010 official estimates adjusted for 2008-2010 infant and child mortality, and old-age mortality. For 1950-2005, due to the lack of adult mortality information and life tables for this period, life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms for males to the South model of the Coale-Demeny Model Life Tables in 1950-1955 and converges over time toward the West model of the Coale-Demeny Model Life Tables and the estimated 2005-2007 life table. For females a similar approach was used assuming that the age pattern of mortality conformed since 1950 to the West model. For each sex, the underlying mortality pattern and implied adult mortality, are consistent with the life table from the 1976-1979 Syrian Follow-up Demographic Survey. For the 2010-2015 period, excess mortality due to the conflict was taken into account.

Chad Derived from estimates of infant and child mortality by assuming that the CD North model life Other HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables Model Life Tables.

Togo Estimated using the South model of the Coale-Demeny Model Life Tables CD South relational Other HIV - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) adjusted estimates of adult mortality (45q15), Adult mortality mortality estimates were derived from (a) recent household deaths data from the 1960 survey, 1970 and 1981 censuses; (b) parental orphanhood from the 1998 DHS, 2000 MICS2 and 2006 MICS3; (c) siblings deaths from the 1998 DHS; (d) implied relationship between child mortality and adult mortality based on the North model of the Coale-Demeny Model Life Tables in 1950-1955 and assumed to converge over time toward the South model of the Coale-Demeny Model Life Tables by the 1990s.

Thailand Based on life tables derived from official estimates of registered deaths and Death registration data Other HIV 1980-2009 enumerated census population by age and sex from 1948 to 2011, adjusted

World Health Organization Page 30

for infant and child mortality and for underregistration of adult deaths.

Tajikistan Based on registered deaths and population by age and sex through 2008, Death registration data vr 1981-1982, 1985- adjusted for underregistration of deaths. 2005

Turkmenistan Based on official estimates of life expectancy available through 2006, Death registration data vr 1981-1982, 1985- adjusted for underregistration of deaths. 2013

Timor-Leste Based on child mortality and adult mortality estimates from the 2009/10 CD West model life wpp - Timor-Leste DHS. Life tables are estimated using the Flexible two- tables dimensional model life table and Lee-Carter method. For 1950-2005, derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the West model of the Coale-Demeny Model Life Tables. Official estimates of life expectancy at birth for the year 2002 were also taken into account.

Tonga Based on: (a) the registered deaths by age and sex from 1957 to 1966 and UN Far Eastern model wpp 1992-2003 from 1982 to 2006 and underlying population by age and sex; and (b) life tables estimates from the 1996 and 2006 censuses by assuming that the age pattern of mortality conforms to the Far Eastern model of the United Nations Model Life Tables. Estimates from the Secretariat of the Pacific Community were also considered.

Trinidad and Based on: (a) registered death by age and sex through 2005 and underlying Death registration data vr 1980-2009 Tobago population by age and sex, and (b) official estimates through 2000.

Tunisia Based on official estimates of life expectancy from 1995 to 2012 from INS Death registration data wpp 1980, 1987-1989, Tunisia. The age pattern of mortality is based on national life table from 1991-2000, 2009, various years adjusted for under-five mortality. 2013

Turkey Based on: (a) adjusted estimates from registered deaths by age and sex from Death registration data wpp 1999-2002, 2004- 1952 to 2006 and for 2009 with underlying population by age and sex; (b) 2013 official estimates for 1989, 2006, 2008 and 2011; and (c) estimates from 1990 to 2010 from the Turkish Institute of Statistics.

China: Province Based on official estimates of life expectancy derived from registered Death registration data wpp - of Taiwan only deaths through 2009.

United Republic Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV 1988 of Tanzania age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates. Estimates of adult mortality were also considered. These were based on: (a) parental orphanhood from the 1978, 1988 and 2002 censuses, and the 1992, 1996, 1999, 2004/05 and 2010 DHS; (b) siblings deaths from the above DHS; (c) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for the period of 1988-2012.

Uganda Derived from estimates of infant and child mortality, and adult mortality by CD North relational High HIV - assuming that the age pattern of mortality conforms to the North model of model for non-HIV the Coale-Demeny Model Life Tables. Adult mortality (45q15) estimates were mortality based on: (a) parental orphanhood from the 1969, 1991, and 2002 censuses, and the 1988/89, 1995, 2001, and 2006 DHS; (b) siblings deaths from the above DHS; (c) intercensal survivorship from successive census age distributions (smoothed and unsmoothed) for the period of 1991-2002. The demographic impact of AIDS has been factored into the mortality estimates.

Ukraine Based on official estimates of life expectancy available through 2013. The Death registration data vr 1981-2012 age pattern of mortality is based on life tables through 2013 from the Human Mortality Database. Both estimates incorporate an adjustment to infant mortality.

Uruguay Based on: (a) registered deaths by age and sex through 2013 and Death registration data vr 1980-2010, 2012- underlying population by age and sex; (b) official estimates from 1964 to 2013 2008; and (c) estimates from the 1963, 1975, 1985, 1996, 2004, and 2011 censuses. The number of deaths was adjusted using the growth-balance method.

United States of Based on official estimates of life expectancy available through 2011. The Death registration data vr 1980-2013 America age pattern of mortality is based on life tables through 2011 from the Human Mortality Database.

World Health Organization Page 31

Uzbekistan Based on official estimates of life expectancy available through 2008, Death registration data vr 1981-2005 adjusted for underregistration of deaths.

Saint Vincent Derived from estimates of infant and child mortality by assuming that the CD West model life wpp 1980-2013 and the age pattern of mortality conforms to the West model of the Coale-Demeny tables Grenadines Model Life Tables. Registered deaths by age and sex through 2009 with underlying population by age and sex were considered.

Venezuela Based on: (a) registered deaths by age and sex from 1950 through 2009 and Death registration data vr 1980-2012 (Bolivarian underlying population by age and sex; (b) estimates from the 1950, 1961, Republic of) 1971, 1981, 1990, 2001 and 2011 censuses; (c) official estimates for 1974, 1975, 1985, 2000-2002 and 2007; and (d) estimates from the 1977 World Fertility Survey and the 1998 Population and Family Survey. The number of deaths was adjusted using the growth-balance method.

Viet Nam Based on life tables derived from age and sex-specific mortality rates from: Death registration data wpp - (a) recent household deaths data from the 1979, 1989, 1990 and 2009 censuses (unadjusted and adjusted for underregistration using the growth- balance and synthetic-extinct generation methods), and from the 2007 Population Change and Family Planning survey; (b) annual deaths for 2009 from the Viet Nam national sample mortality surveillance programme adjusted for infant and child mortality, and for adult death completeness according to capture-recapture survey; (c) direct and indirect estimates based on parental orphanhood and siblings survival from the 1991 Vietnam Life History Survey and 1995/98 Vietnam Longitudinal Survey; and (d) 1979- 1989 intercensal survival estimates adjusted for outflows of refugees and differential completeness of census enumeration. For 1950-1970 life tables were derived from estimates of infant and child mortality by assuming that the age pattern of mortality conforms to the average experienced of the North and West models of the Coale-Demeny Model Life Tables in 1950- 1955 and converged over time toward 1980s life tables. For 1965-1975, excess mortality due to the war was factored in the overall mortality levels based on direct and indirect adult mortality estimates derived from parental orphanhood and siblings survival from the 1991 VHS and 1995/98 VLS, and from the PRIO Battle Deaths Dataset.

Vanuatu Based on: (a) infant and child mortality estimates; (b) parental survivorship UN Far Eastern model wpp - (orphanhood) data by age of respondent from the 1999 census; and (c) the life tables assumption that the age pattern of mortality conforms to the Far Eastern model of the United Nations Model Life Tables.

Samoa Based on: (a) registered deaths by age and sex from 1980 through with the UN Far Eastern model wpp 1980, 1992-1993 underlying population by age and sex, and (b) estimates from the 1999 and life tables 2009 Samoa DHS. The age pattern of mortality was assumed to conform to the Far Eastern model of the United Nations Model Life Tables. Estimates from the 2001, 2006 and 2011 censuses were also considered.

Yemen Estimated using the West model of the Coale-Demeny Model Life Tables CD West relational wpp - and three parameters: (1-2) direct and indirect estimates of infant and child model for non-HIV mortality, and (3) estimates of adult mortality (45q15). Adult mortality mortality estimates were implied by the relationship between child mortality and adult mortality based on the South model of the Coale-Demeny Model Life Tables and assumed to converge over time toward the West model of the Coale-Demeny Model Life Tables by the 1980s. Indirect estimates of adult mortality based on widowhood data from the1979 WFS, as well as parental orphanhood from this survey and the 2004 census were also considered. Official estimates of life expectancy at birth from the Central Statistical Organization of Yemen were also taken into account.

South Africa Derived from estimates of infant and child mortality by assuming that the Death registration data, High HIV 1980-1982, 1984- age pattern of mortality conforms to the Far Eastern model of the United UN Far Eastern model 2013 Nations Model Life Tables. Official estimates from Statistics South Africa life tables for non-HIV and the Actuarial Society of South Africa were also considered. The mortality demographic impact of AIDS has been factored into the mortality estimates.

Zambia Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV - age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

Zimbabwe Derived from estimates of infant and child mortality by assuming that the CD North model life High HIV 1982, 1986, 1990- age pattern of mortality conforms to the North model of the Coale-Demeny tables for non-HIV 1996, 1998, 2002 Model Life Tables. The demographic impact of AIDS has been factored into mortality the mortality estimates.

World Health Organization Page 32

World Health Organization Page 33

Annex B: Data sources and methods for mortality shocks

Natural disasters

Estimated deaths for major natural disasters were obtained from the EM-DAT/CRED International Disaster Database (1). EM-DAT includes epidemics and some man-made disasters that are classified as transport injuries etc, these are excluded from mortality estimates for natural disasters. Since 2000, three major natural disasters that were associated with more than 100 000 deaths have dominated the picture: the Asia tsunami in 2004; the Myanmar cyclone in 2008; and the Haiti earthquake in 2010 (Figure 1). The number of disasters has been declining in the last decade and the number of people affected reached its lowest levels since 2000 in 2012 and 2013. Out of over one million disaster-related deaths during 2000–2014, 61% occurred in Asia where 60% of the global population live, about one in five people reported killed lived in the Americas. Africa (6%) and Oceania (less than 1% of deaths) had much smaller proportions).

Figure 1 Number of people reported killed in natural and technological disasters, 2000–2014

Age-sex distributions were based on a number of studies of earthquake deaths (2, 3) and tsunami deaths (4, 5).

Conflict deaths

Country-specific estimates of war and conflict deaths have been updated for the entire period 1990- 2015 using revised methods together with information on conflict intensity, time trends, and mortality

World Health Organization Page 34

obtained from a number of war mortality databases (described below). These estimates relate to deaths for which the underlying cause (following ICD conventions) was an injury due to war, civil insurrection or organized conflict, whether or not that injury occurred during the time of war or after cessation of hostilities. The estimates include injury deaths resulting from all organized conflicts, including organized terrorist groups, whether or not a national government was involved. They do not include deaths from other causes (such as starvation, infectious disease epidemics, lack of medical intervention for chronic diseases), which may be counterfactually attributable to war or civil conflict.

Methods used previously by WHO for estimation of direct conflict deaths were developed in the early 2000s and applied adjustment factors for under-reporting to estimates of battlefield or conflict deaths from a variety of published and unpublished conflict mortality databases (5-9). Murray et al. (10) summarized the issues with estimation of war deaths, and emphasized the very considerable uncertainty in the original Global Burden of Disease estimates (11) and subsequent WHO estimates for conflict deaths. WHO published estimates for the years 2000 through 2008 used adjustment factors based on conflict intensity developed from an analysis of likely levels of under-reporting (12-15). These adjustment factors ranged from around 3 to higher than 4 in sub-Saharan Africa.

Obermeyer, Murray and Gakidou (16) more recently analyzed data on deaths due to conflict from post- conflict sibling histories collected in the 2002 to 2003 WHO World Health Survey (WHS) program. They used data from 13 countries with more than 5 reported sibling deaths from war injuries in at least one 10-year period to estimate total war deaths for these countries for the period 1955-2002. The authors then compared their estimates of war deaths to the number of war deaths estimated in the UCDP Battle Deaths database (17) to derive an average adjustment factor of 2.96. Garfield and Blore (18) noted that a very small number of war deaths for Georgia resulted in an outlier ratio of 12.0 which heavily influenced the overall ratio of 2.96. They reanalyzed the WHS-derived war deaths dataset excluding Georgia, to obtain an overall revised adjustment factor of 2.21.

The revised WHO country-specific estimates of war and conflict deaths for the period 1990-2015 make use of estimates of direct deaths from three datasets: Battle-Related Deaths (version 5), Non-State Conflict Dataset (UCDP version 2.4), and One-sided Violence Dataset (UCDP version 1.4) from 1989 to 2011 (19-21). Using these three datasets, instead of focusing solely on battle-related deaths, reduces the likelihood that overall direct conflict deaths are underestimated. However, it is likely that a degree of undercounting still occurs in the count-based datasets, and a revised adjustment factor of 1.91 has been applied to the annual battle death main estimates for state-state conflicts. No adjustments were applied to estimated conflict deaths (main estimates) for non-state conflict deaths, and one-sided violence.

The adjustment factor 1.91 is the average of the factor of 2.21 obtained by Garfield and Blore (18) and of a factor of 1.66 derived from comparing total deaths in the UCDP battle deaths dataset with those estimated by the Peace Research Institute Oslo (PRIO) for the years 1989-2008 (22). As shown in the graph below, the PRIO estimates are systematically higher than those of the UCDP.

World Health Organization Page 35

The UCDP dataset is compiled primarily by counting the annual total of combat-related fatalities (national and global) from reports of fatalities in individual violent incidents (battles, clashes, etc.) in each state-based conflict. UCDP uses a variety of sources, including news reports, reports from human rights organizations and nongovernmental organizations, etc. Since it is highly unlikely that all reports of battle deaths will be recorded—particularly in conflicts where outside observers are banned from war zones—this methodology will almost certainly underestimate the actual number of battle deaths. By contrast, the PRIO dataset relies heavily on summary estimates—i.e., expert assessments of overall fatalities. There is no reason to assume that summary estimates will systematically undercount battle deaths as does UCDP’s incident-based estimation method.

Note that the application of a single adjustment factor for all state-state conflicts may result in deaths for specific conflicts being over- or under-estimated. For the following countries, the multiplier was adjusted downwards for low intensity years: Mexico (drug gangs), DR Congo, Columbia, Eritrea/Ethiopia (1990-2000). For these conflict, estimated deaths from other sources suggest that UCDP figures provide reasonable estimates without additional adjustment. For several conflicts where more specific sources of information are available, these have been used to revise estimated deaths:

Iraq (81, 82).

Iraq The conflict death toll in Iraq following the US-led invasion in March 2003 has been the subject of much discussion with estimates for violent deaths to end June 2006 ranging from 47,668 (Iraq Body Count) to 601,027 in a 2006 household survey (). The Iraq Family Health Survey (IFHS), conducted in 2006-2007 by relevant Iraq Government Ministries in collaboration with WHO, provided new evidence on mortality in Iraq for the three years post-invasion (24). Latest counts of reported deaths in Iraq by the Iraq Body Count (25) were compared with conflict deaths for the period 2003-2006 estimated from the Iraq Family Health Survey 2006 (24). This nationally representative

World Health Organization Page 36

survey of 9,345 households included questions on deaths of adult siblings of respondents, and deaths in the household. Sibling deaths were used to estimate adult mortality rates using the Gakidou-King method (26). Calendar year adjustment factors for under-reporting in the Iraq Body Count data ranged from 3.3 (2003) and 3.4 (2004) to 2.3 (2006) and 2.2 (2007). An average adjustment factor of 2.17 was applied to Iraq Body Count data for more recent years to derive a time series of estimated total conflict deaths in Iraq.

Occupied Palestinian Territories. Estimates of Israeli and Palestinian deaths were derived from statistics published by the Office for the Coordination of Humanitarian Affairs (OCHA) - Occupied Palestinian Territory (OPT) (27) and the The Israeli Center for Human Rights in the Occupied Territories (28).

Syria For Syria, excess mortality in 2011 and 2012 due to the conflict was taken into account based on UN estimates of overall conflict deaths by month and age distribution of deaths (29, 30), as well as estimates by various human rights organizations (31, 32).

US, UK and the coalition of the willing. Military deaths in Afghanistan and Iraq were compiled from various official sources and summary tables available on websites.

Deaths due to landmines and unexploded ordinance were estimated separately by country (33). Deaths from terrorist events were separately estimated for many countries without ongoing general conflict using data from the Global Terrorism Database (90) and Terrorism deaths. Terrorism deaths from this database were not added to conflict deaths for Iraq, Pakistan, Afghanistan and a number of African countries to avoid potential double counting.

Legal execution deaths are included in this cause category for GHE2015. Estimated execution deaths were added for the main countries using capital punishment regularly (China, Iran, Iraq, DPR Korea, Saudi Arabia, USA and Yemen), from UN Human Rights Reports, with additional information from Amnesty International reports, Human Rights Watch reports and Wikipedia.

Age-sex distributions for conflict deaths were revised based on available distributions of conflict deaths by age and sex for specific conflicts (10, 16, 24, 25, 27, 35, 36) and on age-patterns for certain country- periods with high conflict deaths included in the WPP2015 life tables (37).

The following tables summarizes and compares various time series of conflict deaths estimates.

Table 5.2. Estimated total global injury deaths (thousands) due to conflict: comparison of various time series and WHO estimates.

Year GBD 1990 WHO IHME-GBD WHO 2013 PRIO IHME-GBD Current (a) 2000-2008 2012 (j) (i) 2013 (k) revision 2016 2015

World Health Organization Page 37

1990 502 - 63 138 94 72 131 2000 656 310 (b) 2000 230 (c) 2000 187 (d) 53 122 90 64 128 2004 182 (e) 95 31 95 2005 238 (f) 26 69 19 42 77 2008 182 (g) 84 35 85 2010 834 18 57 29 48 64 2013 82 31 157 2014 101 194 (a) Estimates and projections by Murray and Lopez (11)

(b) World Health Report 2001 (87) and World report on violence and health (38). (c) World Health Report 2002 (12) (d) Revision for Disease Control Priorities Study (13) (e) Global burden of disease: 2004 update (14) (f) World Health Statistics 2007 (39) (g) WHO estimates of causes of death for year 2008 (15) (h) Sum of main estimates of conflict deaths for state-state, state-nonstate and one-sided conflicts (19-21) (i) Revised WHO estimates for years 1990-2011 (40). (j) IHME Global Burden of Disease Study 2010 (41). (k) IHME Global Burden of Disease Study 2013 (42).

The revised WHO estimates for total conflict deaths (in the final column) are considerably lower than the previous WHO estimates for years 2000-2008 which used the earlier higher adjustment factor for under-reporting, which in turn are lower than the previous estimates and projections in the original Global Burden of Disease (GBD) study (11). The recently estimates for conflict deaths published by IHME in the GBD 2013 study, shown in the rightmost column, are considerable lower than the revised WHO estimates. The IHME estimates are also lower than the main estimate from the UCDC databases for the same year. The IHME methods were based on a regression analysis of available all-cause mortality data for country-years in which battle deaths were reported in various databases. Lozano et al (41) cite (43) for more detailed documentation of their methods. The latter publication does not appear to exist.

World Health Organization Page 38

References

1. CRED. EM-DAT: The CRED International Disaster Database. Belgium, Université Catholique de Louvain, 2013. Available at http://www.emdat.be/disaster-list (accessed 27 September 2013). 2. He H, Oguchi T, Zhou R, Zhang J, Qiao S. Damage and seismic intensity of the 1996 Lijiang earthquake, Vhina: a GIS analysis. Technical report. Tokyo, Center for Spatial Information Science, University of Tokyo, 2001. Available at: http://www.csis.u-tokyo.ac.jp/english/dp/dp.html (accessed 18 January 2008). 3. Naghii MR. Public health impact and medical consequences of earthquakes. Pan American Journal of Public Health, 2005, 18:216–221. 4. Nishikiori N, Abe T, Costa DG, Dharmaratne SD, Kunii O, Moji K. Who died as a result of the tsunami? Risk factors of mortality among internally displaced persons in Sri Lanka: a retrospective cohort analysis. BMC Public Health, 2006, 6:73. 5. Doocy S, Rofi A, Moodie C, Spring E, Bradley S, Burnham G et al. Tsunami mortality in Aceh Province, Indonesia. Bulletin of the World Health Organization, 2007, 85:273–278. 6. Heidelberg Institute on International Conflict Research. Conflict barometer. Department of Political Science, University of Heidelberg, 2012. Available at: http://www.hiik.de/en/konfliktbarometer/. 7. Project Ploughshares. Armed conflicts report. Waterloo, Canada, Project Ploughshares, 2005. Available at: http://www.ploughshares.ca/. 8. Marshall MG, Gurr TR. Peace and conflict 2005: a global survey of armed conflicts, self-determination movements, and democracy. University of Maryland, Center for International Development and Conflict Management, 2005. 9. International Peace Research Institute. UCDP/PRIO Armed Conflict Dataset. Oslo, PRIO, 2009. Available at: http://www.prio.no/CSCW/Datasets/Armed-Conflict/ (accessed 2 November 2009). 10. Murray CJ, King G, Lopez AD, Tomijima N, Krug EG. Armed conflict as a public health problem. British Medical Journal, 2002, 324(7333):346-349. 11. Murray CJL, Lopez AD. The Global Burden of Disease: a comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990 and projected to 2020. Cambridge, Harvard School of Public Health, 1996. 12. World Health Organization. World health report 2002. Reducing risks, promoting healthy life. Geneva, World Health Organization, 2002. 13. Lopez, A.D., Mathers, C.D., Ezzati, M., Murray, C.J.L., & Jamison, D.T. Global burden of disease and risk factors. New York, Oxford University Press, 2006. 14. World Health Organization. The global burden of disease: 2004 update. Geneva, World Health Organization, 2008. 15. World Health Organization. Causes of death 2008: data sources and methods. http://www.who.int/healthinfo/global_burden_disease/cod_2008_sources_methods.pdf . 16. Obermeyer Z, Murray CJL, Gakidou E. Fifty years of violent war deaths from Vietnam to Bosnia: analysis of data from the world health survey programme. British Medical Journal, 2008, 336:1482-6. 17. Lacina B, Gleditsch NP. Monitoring trends in global combat: a new dataset of battle deaths. Eur J Popul, 2005, 21:145-166. 18. Garfield, R, Blore J. Direct Conflict Deaths. Unpublished report prepared on behalf of the Collective Violence Expert Group for the Global Burden of Disease Study, 2009. 19. Uppsala Conflict Data Program. UCDP Battle-Related Deaths Dataset v.5-2015, 1989-2014. Oslo, Uppsala University, 2015. Available at: http://www.pcr.uu.se/research/ucdp/datasets/ucdp_battle- related_deaths_dataset/ (accessed 8 July 2015). 20. Uppsala Conflict Data Program. UCDP Non-State Conflict Dataset v.5-2015, 1989-2014. Oslo, Uppsala University, 2015. Available at: http://www.pcr.uu.se/research/ucdp/datasets/ucdp_non- state_conflict_dataset_/ (accessed 8 July 2015). 21. Uppsala Conflict Data Program. UCDP One-Sided Violence Dataset v.5-2015, 1989-2014. Oslo, Uppsala University, 2015. Available at: http://www.pcr.uu.se/research/ucdp/datasets/ucdp_one- sided_violence_dataset/ (accessed 8 July 2015).

World Health Organization Page 39

22. Human Security Report Project. Estimating Battle Deaths: A Challenging Exercise. Vancouver, BC, 17 September 2012. Available at: http://www.hsrgroup.org/docs/Publications/Additional- Publications/HSRP_Est.Battle_Deaths.pdf (accessed 1 December 2015). 23. Burnham G, Lafta R, Doocy S, Roberts L. Mortality after the 2003 invasion of Iraq: a cross-sectional cluster sample survey. Lancet 2006; 368 (9545): 1421-28. 24. Iraq Family Health Survey Study Group. Violence-Related Mortality in Iraq from 2002 to 2006. N Engl J Med, 2008, NEJMsa0707782. 25. Iraq Body Count. Iraqi deaths from violence 2003–2015. Available at: http://www.iraqbodycount.org/ 26. Gakidou E, King G. Death by survey: estimating adult mortality without selection bias from sibling survival data. Demography 2006; 43: 569-585. 27. Office for the Coordination of Humanitarian Affairs (OCHA) - Occupied Palestinian Territory (OPT). http://www.ochaopt.org 28. B’Tselem – The Israeli Information Center for Human Rights in the Occupied Territories. Statistics on injuries and deaths suffered by both sides in the conflict. http://www.btselem.org/statistics 29. Price M, Klingner and BallPreliminary Statistical Analysis of Documentation of Killings in Syria. UN OHCHR commissioned report, January 2013. Available athttp://www.ohchr.org/Documents/Countries/SY/PreliminaryStatAnalysisKillingsInSyria.pdf (accessed 17 January 2014). 30. UN Secretary General Ban Ki-moon. Statement by the Secretary-General on Fulfilling our Collective Responsibility on Syria. 12 Mar 2015. http://www.un.org/sg/statements/index.asp?nid=8457 31. Syrian Observatory for Human Rights. 7 Feb 2015. http://www.syriahr.com/en/2015/02/about-2-millions- killed-and-wounded-in-47-months-and-it-is-still-not-enough/ 32. Wikipedia. Casualties of the Syrian Civil War. https://en.wikipedia.org/wiki/Casualties_of_the_Syrian_Civil_War (accessed 30 Nov 2015). 33. International Campaign to Ban Landmines. Landmine monitor. Available at http://www.the-monitor.org/ (accessed 30 Aug 2015). 34. National Consortium for the Study of Terrorism and Responses to Terrorism. Global Terrorism Database http://www.start.umd.edu/gtd/ (accessed 30 Aug 2015). 35. Hoeffler A. Dealing with the consequences of violent conflicts in Africa. Background Paper for the African Development Bank, 2008. Available at: http://users.ox.ac.uk/~ball0144/consequences.pdf 36. World Health Organization. European Programme for Intervention Epidemiology Training. Retrospective mortality survey among the internally displaced population, Greater Darfur, Sudan, August 2004. Geneva, World Health Organization, 2004. Available at: http://www.who.int/disasters/repo/14652.pdf 37. UN Population Division (2015). World Population Prospects - the 2015 revision. New York, United Nations. 38. Krug EG, et al. World Report on violence and health. Geneva: World Health Organization, 2002. 39. World Health Organization. World Health Statistics 2007. Geneva: World Health Organization, 2007. 40. World Health Organization 2013. WHO methods and data sources for global causes of death 2000-2011 (Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.3) 41. Lozano R, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet, 2012, 380(9859):2095- 128. 42. Naghavi M, Wang H, Lozano R, et al. Global, regional, and national age–sex specific all-cause and cause- specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet, 2015, 385(9963): 117-171. 43. Murray C, Lopez AD, Wang H. Mortality estimation for national populations: methods and applications. Seattle, University of Washington Press, 2012.

World Health Organization Page 40

Annex C: Estimated completeness of death registration data

Annex Table C: Estimated completeness of death registration data, by country and year, 1985-2015. Albania Argentina Armenia 1.0 - 1 .0 - =--=-= ===" ---=-- 1.0 - =-:.... 0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 -

1H85 2005 2015 198. 1995 2005 2015 1985 19H5 2005 2015

1.0 - --===---- 1 0-

Australia Austria Azerbaijan 1 .0 ------0.8 - 0.8 - 08 - 0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0 0- o o - o o- I I I I 1 I I I 1 I I I 1oes 1005 2005 2015 1085 1005 2005 2015 1095 1905 2005 2015

1.0 - - -=-- 1.0 - 1.0- Belarus Belgium Bosnia and Herzegovina

0.8 - 0.6 - 0.6 - 0.6 - 0.0 - 0.6- 0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 - 0.0- 0.0 - 0.0 - I I I I I I I 1985 1995 2005 2015 1965 1995 2005 2015 1985 1995 2005 2015

Brazil Brunei Darussalam Bulgaria 1.0- 1.0 - 1 .0 - ----,.-...-----.:::::::::::::======­ ===::::::;:;;::::::::::::::::="":> .c::: ­ :=:> 0.8- 0.8 - 08 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 -

1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Key - IHME - WHO (previous) - WHO (current)

World Health Organization Page 41

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

Canada Chile Colombia 1.0 - 1.0 - ---====--- 1.0 - =---=- 0.8 - 0.6 - 0.8 ------

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Costa Rica Croatia Cuba 1.0 - .- 1.0 - --==- s 1.0 - 0.8 - 0.8 - 0.8 - - 0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Czech Republic Denmark Dominican Republic

1.0------1.0 ------=- 0.8 - 0.8 - 0.6 - 0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0 4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

...... 8 - '<::::: - :::;::::> Ecuador Egypt ElSalvador 1.0 - 1.0 - ;:::::---::...- 1.0 -

0. 0.8 - 0.8 - - 0.6 - 0.6 - 0.6 - 0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 - 0.0 - 0.0 - 0.0 - I I I I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

World Health Organization Key -IHME - WHO (previous) - WHO (current) Page 42

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

Estonia Finland France 1.0 - 1 .0 - 1.0 -

0.8 - 0.6 - 0.8 -

0.6- 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Georgia Germany Greece 1.0 - 1.0 - 1 o - ="":::::. .,...--..., 0.8 - <::::::::?' 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0- 0.0 - 0.0 - I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Guatemala Guyana Hungary 1 .0 - 1.0 - -::::::;;;;;:::; ;:;= ;;::::: 1.0 ------'"='" 0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 200 2015 198 1995 2005 2015

Iceland Ireland Israel

1.0 - ======::::::::-__

1.0 - ----==:<:!!!!!!!!!:::=- 1.0 - ...,., 0.8 - 0.6 - 0.8 - 0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Key - IHME - WHO (previous) - WHO (current)

World Health Organization Page 43

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

Italy Japan Kazakhstan

1.0 - 1.0 - 1.0 - ---...... 0.8 - 0.8 - 0.6 -

0.6 - 0.8 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Kyrgyzstan Latvai Lithuania 1 0 - 1.0 - 1.0 - .... 0.8 - 0.8 - 0.8 - 0.6 - 0.5 - 0.6 - --- 0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Lu embourg Maldives Malta 1.0 - - -===:::====:_ 1.0 ------=-""-c'=====- 1.0 ------0.8 - 0.6 - 0.8 -

0.0 - 0.6 - 0.6 - 0.4 - 0.4 - 0.4 - 0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 200 2015 198 1995 2005 2015

Mauritius Mexico Mongolia 1.0 - 1.0 - ======:- 10 - =="""'-:a.- -- _ _ - - 0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.5 -

0.4 - 0.4 - 0.4 - 0.2 - 0.2 - 0.2 - 0.0 - 0.0 - 0.0 - 1 I I I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Key -IHME - WHO (previous) - WHO (current)

World Health Organization Page 44

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

-7"7"'"«"'>0::::...oo:;:<:;::===-- - 1.0 - - - - - 1.0 - 1.0 ------""===- Montenegro Netherlands New Zealand 0 8-

0.6 - 0.6 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0- 0.0 - 0.0 - 1 I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Nicaragua Norway Panama 1.0 - 1.0 ------1.0 ------0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Peru Philippines Poland 1 .0 ------1.0 - 0.8------0.6 - 0.8 - 0.6 - ::?-:-:::::::==:.. - 0.6 - 0.6 - 0.4 - 0.4 - 0.4 -

0.2- 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 200 2015 198 1995 2005 2015

Portugal Puerto Rico Republic of Korea 1.0 - 1.0- --==--======- 1.0 - ;;;;=-='"""==--===--- 0.8 - 0.6 - 0.8 -

0.6 - 0.6 - 0.6-

0.4 - 0.4 - 0.4 - 0.2 - 0.2 - 0.2- 0.0 - 0.0 - 0.0 - 1 I I I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Key -IHME - WHO (previous) - WHO (current)

World Health Organization Page 45

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

Republic of Moldova Romania Russian Federation 1.0 - 1.0 - 1.0 - ...... 0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Serbia Slovakia Slovenia 1.0 - -:::- 1.0 - 1.0 -

08 - 0.6 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0- 0.0 - 0.0 - I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Spain Suriname Sweden

1.0 ------=- 1.0 - 08 - --==------0.8 - 0.8 - 0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

1.0 - - 1.0 - .0 ------Switzerland Tajikislan The fonner Yugoslav Republic of Macedona 1 .._ ____ 0.8 - 0.8 - 0.6 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - 1 I I I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Key -IHME - WHO (previous) - WHO (current)

World Health Organization Page 46

Annex Table C (continued): Estimated completeness of death registration data, by country and year, 1985-2015.

Turkmenistan Ukraine United Kingdom 1.0 - 1.0 - ==-- 1.0 - -===---- 0.8 - 0.8 - 0.8 -

0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 -

0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

United States of America Uruguay Uzbekistan 1.0 - 1.0 - 1.0 -

0.8 - 0.8 - 0.8 ------0.6 - 0.6 - 0.6 -

0.4 - 0.4 - 0.4 - 0.2 - 0.2 - 0.2 -

0.0 - 0.0 - 0.0 - I I I I 1985 1995 2005 2015 1985 1995 2005 2015 1985 1995 2005 2015

Venezuela (Bolivarian Republic of) 1.0 ------

0.6 - -

0.6 -

0.4 -

0.2 -

0.0 - 1 I 1985 1995 2005 2015

Key - IHME - WHO (previous) - WHO (current)

World Health Organization Page 47

Annex D: Estimated completeness of death registration data for most recent year

Country Year Completeness (%) Country Year Completeness (%)

Albania 2009 76.1 Kyrgyzstan 2013 95.8 Argentina 2013 98.7 Latvia 2012 97.5 Armenia 2012 101.1 Lithuania 2013 88.2 Australia 2012 97.7 Luxembourg 2013 91.5 Austria 2014 98.8 Maldives 2011 99.3 Azerbaijan 2011 95.9 Malta 2014 98.9 Bahamas 2012 93.5 Mauritius 2014 97.4 Barbados 2012 75.7 Mexico 2013 100.8 Belarus 2012 97.0 Mongolia 2010 94.5 Belgium 2012 98.5 Montenegro 2010 92.0 Belize 2013 80.2 Netherlands 2013 99.1 Bosnia and Herzegovina 2011 94.7 New Zealand 2011 97.0 Brazil 2013 100.0 Nicaragua 2013 72.0 Brunei Darussalam 2013 100.0 Norway 2013 96.2 Bulgaria 2012 97.8 Panama 2013 93.2 Canada 2011 96.1 Peru 2013 61.7 Chile 2013 100.0 Philippines 2009 86.0 Colombia 2012 76.2 Poland 2013 100.0 Costa Rica 2013 87.2 Portugal 2013 97.7 Croatia 2013 99.1 Puerto Rico 2013 99.8 Cuba 2013 100.0 Republic of Korea 2013 96.9 Czech Republic 2013 100.0 Republic of Moldova 2013 85.7 Denmark 2012 95.3 Romania 2012 99.1 Dominican Republic 2012 54.8 Russian Federation 2011 95.6 Ecuador 2013 83.0 Saint Lucia 2012 81.3 Egypt 2013 95.4 Serbia 2013 89.7 El Salvador 2012 83.6 Slovakia 2014 96.8 Estonia 2012 97.7 Slovenia 2010 98.7 Finland 2013 97.8 Spain 2013 95.4 France 2012 99.0 Suriname 2012 77.6 Georgia 2014 100.0 Sweden 2014 98.8 Germany 2013 100.0 Switzerland 2012 98.5 Greece 2012 98.4 Tajikistan 2005 82.8 Guatemala 2013 92.0 TFYR Macedonia 2010 101.1 Guyana 2011 93.3 Turkmenistan 2013 76.4 Hungary 2013 97.6 Ukraine 2012 95.3 Iceland 2012 97.8 United Kingdom 2013 98.5 Ireland 2012 98.4 United States of America 2013 97.7 Israel 2014 99.5 Uruguay 2013 101.0 Italy 2012 100.0 Uzbekistan 2005 89.5 Japan 2013 99.9 Venezuela 2012 89.4 (Bolivarian Republic of) Kazakhstan 2012 95.0

World Health Organization Page 48

Annex E: Comparison of 45q15 estimatesWPP2015 and GHE2016

Annex Table E: GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Afghanistan Albania

Male 1-emale Male 1-emale

450- . - ::=·-...... • . . ·.

300 •••• ••• 100- •• :250- • • • . •• • • • 50 - I I I I I 201!1196!1 1 995

,,._

Nale Female Male Fen1ale

19& 1990 ""' 2015 1905 2C05 Algeria Angola fOO - ::: ,,._::= ·········... ·····..... ······· -- 100- - ::!00- ••. •• ••• 0 .

1085 1005 2.)151985 2005 2015 ,., , 2005 2015 "'" '''" .. I I 2COS

Ant1gua and Barbuda Argentina Male Female Male Femael

200- 200- - :=: ·-··-- 160 - ·-- 140- .• •. 0 0 00 120- • •• 100- ···... 100- I I I I I I I I I I I 1985 1990 2005 20151985 '9:)5 2005 2015 1005 1995 2C05 20151385 1995 2005 2015

ale 1-emale World Health OrganizationArmenia M Page 49

A u s t r 140- a l i a M a l e

1 - e m a l e

120- ·.• • 100- • •

:= I I I I I I I I 40 -I k 19&5 1990 2005 2015 19&5 " 990 2000 2015 1905 199!1 2C05 201!1 19&5 1995 2005 2015

Austria AzerbaiJan

150- ' 260_-: •

Male Female 00- • Male Female

100 - · 1SO- • •

100- 50- I 1 I I I • 1 2015 1095 2005 4015185 1QQS 2005 2015 085 1005 2005 20151985 '0;15 2005 '''" ------series - WHO • • ·· WHO. HIV-and shock + WPP

World Health Organization Page 50

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Bahamas Bahrain Male Female Ma' e female

2- 2 - ·· ... "12"0--

100- eo- eo -

1005 2005 2015 Ui8!i 1005 2005 2015 1t5 2005 201 51065 11105 2005 2J15

Bangladesh Barbados

Mako Ma'a Female

-- 1>0- ·········· ····· .

100- .. . ::JL ·.. 1- , ... 1<:;8:> 199) 200:> 201:, IS8t 199:> lOOt 201!1 2l1:>

an1s Belgium Female Ma'e Female

100 - H0- :=: . 120 - 2- 100- 2>l-

100 - eo- tl-

1085 1005 2005 20151G85 1005 2005 2016 1 E5 I GQS 2005 20151985 1.. 5 2005 2)15

Bel1 ze Benin Mako Female Ma'a Female

250- ·····. ·.. ··· 350 - 2 - ..· • 300- ...... _ 151)- ,,.,_ ..··...... / . 1>l-

1 , """ 111\/o '"" lU1t>1 :>

Bhutan Bolivia (Piurinational State of ) Male Female Ma'e Female

35<)-

:D:= - •

2J(j - •

2 - . I I I I I I I I 1085 1905 2006 2015 1G86 1006 2005 2016 1;>t5 1G06 2006 20161006 1.. 6 2005 2J15

series - WHO · · ·· WHO, HIV-and shock • WPP

World Health Organization Page 51

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Bosnia and Herzegovina Botswana Male Female Male Ferrale

400- 200- • 0. • 300- _A == = 200- •.. .• . •• • . • •• • .•0. 100 - ···········..... too- I ?(t1'11M5 701!\ 1 1'ifi 199S

Brazil Brunei Darussalam Male Female '-!ale Ferrale

:l>O- 250 - . • • • . .••• u n - •

200- • . •.. 100160 - 1 20- 150- - . 100- M- • I I I I I I 196 199 2000 199!:i 201:i 196 199:i 200:i 201:iH8G 2:005

Bulgaria Burkina Faso Male Female '-!ale ferrale

150- 2«> - ••• . · 100- :: ...... ···· I I I I I I I I I I 1965 1995 2005 1985 1995 2005 201 5 H85 1995 2005 4015

Burundi Cabo Verde Mala Female '-!ale Ferrale

250 -

200- •.• .. .. ISO- •,• • •, ·· ··.... ··.. 250- . :: 100- ••• • •. I 10'>5 2005 2C15t08S 10'>5 2015 1!)85 2005 201 51)85 1005 2005 2015 """ "'"

Cambodia Cameroon Mal Female '-!ale Ferrale

400-

- - ;:= )( )- • 35< > . 350- ...... · 250- ••• • - • . . lOO- . . • • • • • 200- • . • • .. ' • · -. 150- • .. •• • 2:,0- ...

,., lU1 19\1) :ltl lHIS ,,., :.:ot """

series -WHO ·· ·· WHO, HIV-and shock • WPP

World Health Organization Page 52

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Canada Cenlral African Repul>ic M Female Male Female . ,. , ,

ale == ...... \ '"'- /. 350- •• . 120- • • .·. 100 - • •.••• •

300- • • . ••••• . . • ••••••

:= I I I I 250- 1 1965 zco 201:5191!!5 1990 2005 201!5 199S 2005 199!5 2005 201 ""' ""'

Chad Ch1le Male Female Male female

2<0 - 150- - . • .. 350- ... 100 - • ..

300- 50- - I 1985 1995 2COS 201519€5 199S 2005 2015 1985 199S 2005 i!:01S 198S 1995 2005 ''"'

China China: Province of Taiwan only Male Female Male Female ::= 140- 1:0:0-

100 - 1CO- =:: · · .. .,- 80- · 60- 1 19815 1... 2011519E5 1... 2006 2015 1086 2006 "'"

Color11bia Comoms Mule Fcmab Mole Fcmolc

250 - . . 200- .. ••·._..·.... 150- ­ 100 - ·

1 .5 1005 200S 201519ES ,... 2005 2015 1085 ,.. , 2005 zo1stgss 19{15 2005 2015

Congo Costa Rica Male Femal9 Malo Female

140 - A A ' 7600- "1,., 1:::0- ...- "'f"#' 500- lCO-

::= · ··-·-"··... 80- 200- ·· . . • . ••• I I I I ?01!iiNF5 ?015 ?00 ""' '"'"

World Health Organizationse ries - WHO ·· · · WHO, HIV-and shock • WPP Page 53

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

C te d'lvoire Croatia Male Female Male Female

200- 4 - .2CO- ,

400- / • .··"· ··. 350- . ... . • • • • •• ...... 150 - •

...... 1(0- . 300- • • .•.. so - I 1965 zco 201:5191!!5 1990 201!5 2005 ""' ""' '""'

Cuba Cyprus Male Female Male female

1(0 -

:::= so- 100 - 60- w- 40 -

1985 1995 2COS 201519€5 199S 2005 2015 1985 2005 i!:01S 198S 1995 2005 '""' ''"'

Czech Repubilc Democratic People's Republic of Korea Male Female Male Female

200 -

150-

100-

50- I 1 I I I 19815 1... 2011519E5 1 . 2006 2015 ------

Oen1oaatic Republic of the Congo Denmar k Mule Fcmab Mole Fcmolc

1W - 400- HO - 350- ···········... --- I:iO- 300- • • • . • ...•• 100- 250- • • • • • • ••••••• •• 60-

00 -

1 .5 1005 200S 201519ES ,. 2005 2015 1985 ,. . , 2005 zo1stgss 19{15 2005 2015 ..

Djibouti Dominican Republic Male Femal9 Malo Female

350- ... 250- ... ·. . 300- ..• • •••• 2(0 - :...•': . :.

2:i

series - WHO ·· · · WHO, HIV-and shock • WPP

World Health Organization Page 54

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Ecuador Egypt Male Female Male Female 25()- , · - - -

160 - 100- ,.

=-.. I I I I 2011911!J 201l!:ftl!l ""'" """ "'"

El Salvador Equalorial Gu1 nea

Male

400- 400- ·-- 300 - .•• • • 350- ·. 300- •• ••.• . • • • ••. • 200- ?SO- 100- I I I I '9 5 1995 2005 20151985 1995 2(05 2015 19!15 1995 2005 20151985 19f5 2005 015

Eritrea Estona Male Female Malo Female

400- :m- =A= 200- ·•• • , :=JOO- :- " - . 100-

200 -1 I I .. I

.,., _Mole ,

Female Male

2015193.5 "'" 2005 015 ·QM 1 5 2005 20151985 1995 2COS 2015 19!15 1995 2005

Fiji Female Ethiopia

260- ..,._ ...... • 300- • • • . • . . . 200- 300- • •• • • • • ••

25<)- ••• • • •• • • •

200- • • • '. 15()-

1. .5 2005 2()151085 1"'5 2COS 20Hi 1085 1.. 5 2005 2015101)5 200

Finland 100- Frarc:e Male Female M•le

World Health Organization ·-- Page 55

Female

50 - 50- I 1995 2005 20151985 1995 2COS 20Hi 1965 1995 2005 20151985 19 5 2005 01·!

series -WHO ····WHO, HIV-and shock • WPP

World Health Organization Page 56

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Gabon Gambia

Male Female Ma' e female

3>0-

300 - --

:: 250- · . · - ...... '"'- f i

200-

1005 2005 2015 Ui8!i 1005 2005 2015 1t5 2005 201 51065 11105 2005 2J15

Georgia Genmany Mako Ma'a Female

2>0- 160 - uo - 2 - ·.. 120- • 100- eo- 1J 201:,0158:, ""' 2l1:>

Ghana Greece

Female Ma'e Female

3J

Grenada Guatemala

Mako Female Ma'a Female

250- 300- 2 - - ·. · . ·... 15<)- ::

1&0-

, , , , lU1t>1 :> .. """ .. '""

Guinea Guinea-Bissau Male Female Ma'e Female

35<)- \ 3>0-

3J

25 2 - ---· · ·. .· .·. 200-

1085 1005 2006 2015 1G86 1006 2005 2016 1;>t5 1G06 2006 201 61006 1.. 6 2005 2J15

series - WHO · · ·· WHO, HIV-and shock • WPP World Health Organization Page 57

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Guyana Ha1ti

30 ..... 700- J

Male Female Male Female

2:>0- .• • ··

400- 200- := 300- • • . .. • • • • ··········· 200- ·..• • • .•. .• . • . - tso- I 1 ·g:s 2011911!J 20119tl!l ""'" """ "'"

Honduras Hungary Male

300- 7 - :=- :..=._ A 200-

100- I I 100150- ·...... 160- •·• • • · •• ··• · I I '9 5 1995 2005 20151985 1995 2(05 2015 19!15 1995 2005 20151985 19f5 2005 015

Iceland India

=20·0-....

•• 1000-- oo- Male Female Malo Female 60- • .

40 - 150- I I I 1995 2005 20151985 1995 2COS 2015 19!15 1995 2005 2015193.5 2005 015 "'"

Indonesia Iran ! slamic Republc of)

Mole Female Male Female 400- • ?40- 220- JOO - • 0 • • 200- • •

180-

160-

1. .5 200S 2()151085 10<:S 2COS 20Hi

50- Iraq

Male 200- •••. • . World Health Organization 2 Page 58 JOO- \l}V

Ire and 1SO- • '·• • • •• • Female M•le Female 100- I

:120- := 100- 80-

60- ·ga 1995 200S 20151985 199S 2COS 20Hi 1965 1995 2005 20151985 19 5 2005 01· !

series -WHO · · · ·WHO, HIV-and shock + WPP

World Health Organization Page 59

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Israel Italy Male Female Male Female

120 - •. ••• . ::= · · 110040 - •. • •• . := . := • , '• . ----- 40- · 40- · I I I I I I I I I 2011911!J 201l!:ftl!l ""'" """ "'"

Jamaica Japan Male

120 - =- 100-

1 - •• "'- . $0- 100- 40- I I I I I '9 5 1995 2005 20151985 1995 2(05 2015 19!15 1995 2005 20151985 19f5 2005 015

Jordan Kazakhstan Male Female Malo Female

100- I I I I I 19!15 1995 2005 2015193.5 2005 015 "'"

Kenya Kiribati

<3500- Moe F emale

400- ·._ .. Male Female /\ 300- 250- . . .. • . . • • . • • . • .. • • . • • A• 200 - . ..••• •• •••••.•• • • tso- I 1. .5 2005 2()151085 1"'5 2COS 20Hi 1085 1.. 5 2005 2015101)5 200

Kuwait Kyrgyzstan

Male Female

M•le Female =- 600- 100 - . 500- ..,._ World300 - Health Organization Page 60 L

300 -

2 - 1995 2005 20151985 191l5 2COS 150-

100-

20Hi 1965 1995 2005 20151985 19 5 2005 01· !

series -WHO · · · ·WHO, HIV-and shock + WPP

World Health Organization Page 61

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Lao People's Democratic Republic Latvia Male Female Male Female

1{)85 1005 "''' 20151085 1006 2005 :101f 1!)86 100E 20)5 20161!:85 1096 2005 :1016

Lebanon Lesotho Male Female Male Female

>X- 2)() - ::= · . 4'X- 1:>(.)- 0 ... ::= 3ClC- • •. , ., , - - . ········ ... 20C- so - • • I I I I I I I I I I I I 1985 1995 2005 20151985 1995 2005 2015 1985 1995 20)5 2o1s1ess 1995 2005 2015

liberia Libya Female Male Female

::= _ j A, :B C:-: 10C- . • ::=.V- \--J . A •. A . 14C- 25<- •• • •• • . ...·• ·.-. q-!:':':(..-·-- 12C-

2'JfJ-

19&5 199S Z005 20151985 1995 2005 01! 19&5 1995 20)5 2015tE85 1995 2005 :1015

Li: uania Luxembourg Male rem!.le Male remele

35<- JJ

2 J

15< -

1J

Madagascar Mal91wi Female Male Female

3>0 -

3- ·• ,._ . - ...... •

19&:5 201:51985 ,.., 2000 ,., ,.., 2005 19&5 2)()- I 201:5H:85 21JC- ········· '

series WHO · · ·· WHO, HIV-and shock • WPP World Health Organization Page 62

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Ma'aysia Madives Male Fem.:!le ale Female

2)()- 300- ·.·.. 15< - . :,,.,_ := 11)0 - ••

!:0- I I 1085 005 1095 2005 2015 1085 :i005 20151,8$ 1095 2Co05 2015

Mah Malta Femal

=:: ::= 25< - -10-

,,., 199 2005 ,,.., '''"

Mauritania Mauritius Male Female N'ale Female

200 - •. ?4::=r• - ·.-.. ?00- 22C'- '• • . .. 100- · '"'- ... 100 - 13£1-1 1085 ,.. ;;oos 1095 2005 4:015 1015 4005 20151,85 1.. 5 2(()5 2015

Mexico Micronesia (Federated States of)

1985 1!195 4005 2015Hi85 ,,.., 2005 2015 1985 2005 20151 85 1995 2015

Mongolia Montenegro Male femele ale Female ::= '-. 140- - - ::= 120 - ::= 100- 1SC- _./ - •o- I I I I 1085 1... :mos 2005 :'t:015 1085 1005 :mos 2015 Ui85 1.. 5 2(()5 2015

series -WHO · · ··WHO, HIV-and shock • WPP

World Health Organization Page 63

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Morocco Mozambique Male Female Male Female -· 400- • •

35()-

300-

260 -

""'" 2011911!J """ "'" 201l!:ftl!l

Myanmar Namibia Male

•

==.-. L· == -. /\..'v J\ 200 - • • • • .. • . 200 - . ·•·• • • ··• ·.•• •. . ---- ... • ··... ····· I I I I '9 5 1995 2005 20151985 1995 2(05 2015 19!15 1995 2005 20151985 19f5 2005 015

Nepal Netherlerds Male

150 - I I I ' 1995 2005 20151985 1995 2COS 2015 19!15 1995 2005 2015193.5 2005 015 "'"

New Zealand Nicaragua Mole Female Male Female

:1:2=0- = =:= ' ; · . ·.- -.. := 100- ·- ;"":""" 60- I I I I I I I I ·o E- 15 2005 20151085 1Q0S 2COS 20Hi ,. . , 1... 2005 2015101)5 200

Niger Nigeria

Male Female M•le Female

:¥; )-

=·······...... ······... ·.._ ·..

200 - •••

1995 2005 20151985 1995 2COS 20Hi 1965 1995 2005 20151985 19 5 2005 01· !

series -WHO · · · ·WHO, HIV-and shock + WPP

World Health Organization Page 64

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Norway Occupted Palestine Territories Male Female Male Female

1200 -

100 -

80 -

60 -

zco 201:5191!!5 1990 201!5 2005 199!5 2005 201 ""' ""' '"'"

Oman Pakistan Male Female Male female

Panama Papua New Guinea Male Female Male Female

:=: .A .•teo- o- 190- 140- •. >«>- ····..• . ·,·· 120 - - · ···· . 100 -

so- 2!0- . • •• • 0 ••

19815 1... 2011519E5 1... 2006 2015 2006 2016 ISS' 1995 2005 .... "'"

PaJC:tguay Peru Mule Fcmab Mole Fcmolc

200 -

1SO - - ·. :..·. 150 - • •. ::1&>-- •••: •••••••• ••••• •• ••• ·. ···.. 1(0 - • •• • • 0 120- I ,.., 1 .5 200S 201519ES ..., 2005 2015 1085 ..., 2005 zo1stgss 19{15 2005 2015

Philippines Poland Male Femal9 Malo Female

300- 2!0-

2SO - zoo-

IW- 200-

1(0- 1"' - 1 I I I ?00 '"'" ------

series - WHO ·· · · WHO, HIV-and shock • WPP

World Health Organization Page 65

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Portugal Puerto Rico Male Female Male Female

25<>- 200 -

150-

100-

00- 1 I I I I 1 ·g:s 2011911!J 20119tl!l ""'" "'"

Qatar Republic of Korea

?5<1-· 200-

15<>-

100-

!0- I I I I I I I I '9 5 1995 2005 20151985 1995 2(05 2015 19!15 1995 2005 20151985 19f5 2005 015

Republic of Moldova Romania Male Female Malo Female

25<>- . /\._ :250 -: , • . , ·; \.. 200 -...... ·

200- 15<>- 150- 100 - 100- I I I 1995 2005 20151985 11X>5 2COS 2015 19!15 11X>5 2005 2015193.5 2005 015 "'"

Ru5sian Federa1ion Rwanda Mole Female Male Fcmt lc

1600000-- 400 - • • • • 600- • • - + 200- ' 400- • ..••.• • •. •. •. • .

200- 100- I ·o E- 1. .5 200S 2()151085 101:S 2COS 20Hi 1-65 1WS 2005 2015 Hi85 1005 2005

Saint Lucia Saint Vincent and the Grenadines Male Female M•le Female

200- . . . 1&0:·=·.· - • •• • • 180- • • • • •...• .• ·.I.. 0 160 - ••• • 100- 140- 140- 1£0- - --

1995 200S 20151985 199S 2COS 20Hi 1965 1995 2005 20151985 19 5 2005 01·!

series -WHO · · · ·WHO, HIV-and shock + WPP

World Health Organization Page 66

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Samoa Sao Tome and Principe Male emale P..lale

30 250- ::·=.. . . 220- ·. . 200- 1200 - . 100 - m- . 160- 1995 201.51!ii8S , , 2015 1 985 , :101.51985 2005 2015 "'" .. '"'' ... "'" '''"

Saudi Arabia Senegal Male Female P..lele remate

300- 160- := ············· ···· 1JI O- 120- :::

100- w- I I 150- I I I ,,., 1990 20CO 2:01$ 168 199 2 )00 2010 1 9M 1 990 lCO' 201!19M 1 99 2005 2:01'

Serbia Seychelles M•la

200-

160- • 200-

150- 100- 100- --.-

1085 1995 2015 HiSS 1995 2)05 2015 1 085 1995 2C05 201!1085 ,... 2005 2015 I I I I I I

Sierra Leone Sln apore Male F motle

1"'0 0-- 120- 5w5o0- ••• ·...... 500- •. . . 100 - 450- ·• · .. Ill-

·.. . ·. 0 60- 400- . . 0 • • • ... .. 40- ,..,, lUbllits!l ,,., ll'll) """

Slovakia Slovenia Male Female fl.lale Femate

250- -

200 - ::150 - 150- 100- 100- so- I I l I I I 1005 1995 20C5 2015Hi8S 1995 2)05 2015 1 985 ""' 2COS 201!1985 l OSS 2Co05 2015

series -WHO ··· · WHO, HIV-and shock • WPP

World Health Organization Page 67

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Solon-en Islands Somal1 a Male Female Male Ferrale

:lOO - 54500- 250- - :: ············ ..... 1- 1 ?(t1'11M5 701!\ 199S

South Africa South Sudan

- .. :: ...... ::= ········...

Male Female Male Ferrale

200- • •• • •••..• • ' • , 450 -

JOQ- . • •

2!i0- l 196 199 2000 199!:i 201:i 196 199:i 200:i 201:iH8G 2005

Spain SriLanka Male Female Male ferrale

140- ··· •• wo- 120 - • • •• l A 250- ••.," ·v.\_ 100 - 200- · • · so- 150-

60- 100 - 40- I I I I I I I I I I I 1965 1995 2005 1985 1995 2005 201 5 H85 1995 200S 4015

Sudan Suriname Mala Female Male Ferrale

350- JOO- • .. • •••• •• · - 0 ······ · =200·- ·..... •••• ?50- • • •• • • ·• ·

- ·. 1065 10'>5 2005 2C15t08S 10'>5 2015 1!)85 "'" 2005 20151)85 1005 2005 2015

Swaziland Sweden Mal Female Male Ferrale

:=::p ...... _/'(...... ::::= World Health Organization Page 68

I I 11.1> lU1 ,,., """

series -WHO ·· ·· WHO, HIV-and shock • WPP

World Health Organization Page 69

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Switzerland Syrian Arab Republic Female Male Female

1 - •• 0-· lJO - • •OlWJ- - + j so- 200- ,.... so- ,,- ·oo - •. • 1!)86 1!)(}6 2006 20161085 '0)5 2015 1(85 2005 :2011: 1!)85 1005 2016 "'"

Tajikrstan Tl1ailand Mole Fem3le M3le Fell'l31e ,,_ _ • .. 2())- ····.•• :: . ····· 2- J·· ·s.J - 150- '() - ··..... IJO - I I I I I I I I 1935 1995 2005 20151985 19;15 21)()5 2015 1 65 1995 2005 201!: 1985 1995 2015

The former Yugoslav Republic of Macedonia Ttmor-Lesta Male t-emare Male

ISO -

140 - ' -- 120- •3o0oJ-

- 2()) - 80- I I I I 1905 1995 zoo 2015196:; 201:1 1995 2005 201!: 1985 1995 2015 ""'

Togo Tonga Female Meale Female

20) - --...... __ _ - J>O- 'M- ...,_ 3:>0- • • • • • • • • • • . • A '41)- \_ '"'- -.-/.... 120-

······ '0) -

1035 2005 20151985 19 5 2t)()5 2015 1995 2005 20119&5 1995 20)5 21)15 ""'

Trinidad and Tobago Tunisia

250- ········· 2 -

150- ·... I I I I 19& 2000 201!1198!5 19 5 2000 2015 ::2005 20119& 2•)1!1 ""'

series - WHO · · · ·WHO, HIV-and shock + WPP

World Health Organization Page 70

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Turkey Turkmenistan Male Fem.:!le ale Female

300-

250-

£00-

150- I 1085 :i005 20151,8$ 1095 2CoOS 2015

Uganda Ukraine Male Femal

:30!'1:·=·"".- •• 200 - ..•

·::=- ...... 200- 3:><- "········• . . ··...... •. ... ···• •• • • 1&0 - ,,., _ 100- I I 198> 1,.., 199 2005 1911!1 1''" 11l

United Arab Emirates Unrted Kingdom Male Female N'ale Female

140 - 16<- uo- .

100- :::=1X• - eo •o- --...... ,.,_ co - -

1985 1005 ;;oos 1995 2005 4:015 1015 4005 20151,85 1.. 5 2(()5 "'"

United Republic of Tanzania United States of America

Female 45()- · . 180- · 160- • •• 0 ••••• Male 140- • • • . • •• . 120-

4X' - •

35(J- . - .. • JJt.l - 0 ..... •• •

Z\ 100- ·. ,,.,_ eo ­ I 1985 1W5 4005 2015Hi85 1""5 2005 2015 19115 2005

Uruguay Uzbekistan Male femele ale Female

2:><-

15(1 - • • 200- ·. . 150- 1J()-

1085 1005 :mos 2015 "85 1095 2005 :'t:015 1085 1005 :mos 2015 Ui85 ,... (()5 2015

I I I I • 2

World Health Organization Page 71

series -WHO ····WHO, HIV-and shock + WPP

World Health Organization Page 72

Annex Table E (continued): GH£2016 and WPP2015 estimates of 45q15, by country, sex and year, 1985-2015.

Vanuatu Vanezuela (Boilvarian Republic of) Male emale P,.lale

260- 200- :: • • • ••

200-

1:;() - - 100- .; · .. 100- I 1965 199 201;'i 1585 1995 2:>05 2015 19&5 1995 2(05 1995 2«lS 2015

Viet Nam Yemen Male emale P,.lale 1-emate

200- "------.

300-

150-

2SO- 100-

700-

2(()5 1965 1005 2XS 20151505 1995 2)05 2015 1905 1995 2COS 2011905 19SS 2015

Zambia ZimbabY/e Male Female Pl.lale Femate 800- • •

ooo- 600 - -1 .\ ..,._ ...... 400- . ' • , ····· 200- 200- ········· ········ I I I I I ,.,., 1005 2XS 20151GSS ,... 2)05 2015 1085 ,..,. 2C()S 201E 1085 1()!)5 2(<)5 201!

series -WHO ··· · WHO, HIV-and shock • WPP

World Health Organization Page 73