The material presented here was last updated in September 2015. Further updates will be made available on this website

The aim, where possible, was to get the mortality data at 5-yearly intervals from 1950 to at least 2015 (and many countries now have data available for 2013). This was mainly provided by the World Health Organization, from the official data provided to WHO by its member states. The population data for single years from 1950 to 2015 comprised the 2015 Revision of the United Nations World Population Prospects.

Where the results from a number of different countries were to be combined (to obtain results for 'All developed countries', 'EU28 (European Union)', 'EU15 (European Union)', 'EU13recent (European Union)', 'Caucasia' or 'Central Asia') then in a few instances evidence about some particular disease was missing from some of the contributing populations. In this case the rates in the populations for which it was available were applied to the populations for which it was not. (For example, in a few populations Hodgkin's disease was not separated from non-Hodgkin's lymphoma.) So, where numbers of deaths from some disease are given for a group of populations, that number is intended to apply to the whole group, including the few populations that contributed no data on that disease.

General notes:

Minor differences between this and previous editions are due to the use of slightly different population estimates and use of actual (instead of estimated) national mortality data for some years.

To each of the large totals, 0.2% was added (see Peto, Lopez et al, 1992). Here and elsewhere, the 'rounding' of certain numbers may cause minor apparent discrepancies in the least significant digits that are displayed.

For five former Yugoslav republics (Croatia, FYR Macedonia, Montenegro, Serbia and Slovenia), where historical trends may be unreliable or are unavailable, pages showing trends are omitted.

For a few populations, special methods had to be adopted to deal with missing data (in addition to those described by Peto, Lopez et al, 1992):

Armenia, 2005: The cause-specific, age-specific 2006 mortality rates were applied to the year 2005 population. Note, data for this country are included in data for Caucasia and not shown separately.

Australia, 2005: The cause-specific, age-specific average mortality rates for 2004 & 2006 were applied to the year 2005 population, and the numbers of deaths in each group were then scaled to give known year 2005 total mortality.

Azerbaijan, 2005, 2010: The cause-specific, age-specific average mortality rates for 2003 & 2007 were applied to the year 2005 population; and 2007 mortality rates to 2010 population. Note, data for this country are included in data for Caucasia and not shown separately.

Belarus, 2005, 2010: The cause-specific, age-specific average mortality rates for 2003 & 2007 were applied to the year 2005 population, and the numbers of deaths in each group were then scaled to give known year 2005 total mortality. The 2009 mortality rates were applied for the year 2010 population to estimate the 2010 mortality.

Belgium, 2000: The cause-specific, age-specific 1999 mortality rates were applied to the year 2000 population, and the numbers of deaths in each group were then scaled to give known year 2000 total mortality.

Germany, 1955, 1960, 1965, 1970, 1973-1989: By addition of separate estimates for the former German Democratic Republic (GDR) and the Federal Republic of Germany (within its pre-1990 territories) (FRG). For 1955, 1960, 1965 and 1970 the cause-specific, age-specific death rates for the former FRG were applied to the population of the former GDR for those years and the numbers of deaths in each group were then scaled to give the known total mortality for those years in the former GDR.

Greece, 1955, 1960: For 1955 and 1960, cause-specific, age-specific rates were estimated by extrapolating backwards (from 1975 and 1965 to give 1955, 1970 and 1965 to give 1960). These rates were then applied to the populations for Greece for the relevant years, and scaled to give known total mortality for these years.

Italy, 2005: The cause-specific, age-specific average mortality rates for 2003 & 2007 were applied to the year 2005 population, and the numbers of deaths in each group were then scaled to give known year 2005 total mortality.

Luxembourg, 1955, 1960, 1965: Belgian cause-specific, age-specific death rates were applied to the Luxembourg populations for the relevant years, and scaled to give known total mortality for these years.

Poland, 1955: Overall mortality from smoking was estimated by smoothed extrapolation back to 1955 of the age-specific smoking-attributed mortality rates for 1965 and 1975 (and mortality from particular diseases in 1955 described as unknown).

Portugal, 2005: The cause-specific, age-specific average mortality rates for 2003 & 2007 were applied to the year 2005 population, and the numbers of deaths in each group were then scaled to give known year 2005 total mortality.

Turkmenistan, 2000: The cause-specific, age-specific 1998 mortality rates were applied to the year 2000 population. Note, data for this country are included in data for Central Asia and not shown separately.