—from Epidemiological Bulletin, Vol. 23 No. 4, Diciembre 2002—
From Basic Data to Composite Indices:
A Re-examination of Mortality Analysis
The importance of mortality data was recognized long before a
much broader concept of public health had been developed. The first example
of systematic death counts dates back to the black plague in England in the
1500s and by the mid-1600s, John Graunt had instituted a systematic collection
of the numbers and causes of death and applied the first life-table methodology
to these data.(1) For most of the following centuries, the characterization
of the health of populations was based on mortality and survival. Although the
description of health has evolved in time to include additional dimensions,
deaths remain among the most important public health events to be measured and
analyzed.
The main source of data on deaths is civil registration. Although in most countries
of the world there exist institutional, legal and technical settings to perform
the recording of vital events, underregistration and the variable quality of
the information on these events are pervasive. In the Region of the Americas,
underregistration was estimated to vary between 0.5% in the United States to
92.1% in Haiti around 1997.(2) The percentage of ill-defined causes of death,
an indicator of the quality of mortality data, also varied from 0.5% in Cuba
to 44.7% in Haiti.(2) Other sources of mortality data available for use, depending
on the characteristics of the country, include health services, cemeteries,
and even police records and the press.
Mortality data are the basis for a wide range of indicators of varying complexity.
They not only represent tools to assess the risk of dying in a population and
the impact of diseases on health, but also the severity of diseases and survival
experienced by the population. As such, they are an essential element in health
situation analysis, public health surveillance, program and policy planning
and evaluation.
Because of the importance in public health to understand and utilize mortality
indicators, over the last 23 years, articles have been regularly published in
the Epidemiological Bulletin on characteristics of mortality data, techniques
and tools used for mortality analysis, and changes in the countries’ mortality
profiles in the Region. There is now renewed interest in this topic, particularly
in the context of monitoring the United Nations’ Millennium Development
Goals.(3)
Over the next few issues, the Epidemiological Bulletin will revisit this
topic in a systematic way and provide articles on both the traditional measures
of death as well as the latest developments in this field. These articles will
be designed, not only to enlighten the technical and practical aspects of mortality
measures, but also to prepare readers for the application to public health challenges,
hopefully promoting two essential public health functions — health situation
analysis and surveillance in public health. Following are the seven general
topics that will be addressed in future issues of the Epidemiological Bulletin.
(1) International Classification of Diseases
There exist important issues regarding the level and quality of the registration
of deaths that impact on the usefulness of mortality statistics. To address
them, themes related to the International Classification of Diseases, such as
the registration of vital events and methods to adjust for under-registration,
death certification and coding, and tabulation of mortality data will be presented.
This issue of the Epidemiological Bulletin includes a first article in the series
that deals with the creation of short lists for the tabulation of mortality
and morbidity data.
(2) Mortality rates, ratios and proportions
One of the main uses of mortality rates is to indicate the overall and absolute
magnitude of mortality in a population. General mortality rates present an array
of options for comparative analysis. These options include population-specific
rates such as sex- or age-specific, maternal or infant mortality rates, and
cause-specific mortality rates. Mortality data may also be used to measure the
severity of diseases and the impact on the population of specific causes of
death. In summary, they give substance to the mortality profile of a population.
Detailed presentation of mortality rates, proportional mortality, and case-fatality
rates will be included in the series of articles.
(3) Survival and life expectancy
Studying mortality implies not only looking at what causes people to die,
but also when and where they die. One of the major objectives of public health
is to postpone preventable death, reducing the burden of avoidable and premature
mortality. Mortality data are used to analyze the survival of a population and
to calculate life tables that allow estimation of the life expectancy of individuals
and population groups. Survival analysis and life tables provide a probabilistic
picture of time to death that can be used to compare populations with different
exposures to mortality risk factors. The concept of survival is also used in
longitudinal epidemiological studies and clinical trials, as a way to express
prognosis. Another use of premature and avoidable mortality indicators is to
evaluate and improve health services.
(4) Period, cohort and age effects
The study of mortality patterns includes examining the evolution of indicators
over time. This requires an understanding of the impact that risk factors associated
with different variables may have on mortality. A future article will address
the impact of events or processes that occur during particular dates or calendar
periods, called the period effect; that of characteristics of a birth cohort
over time, called the cohort effect; and the impact of aging of a population
on mortality, called the age effect.
(5) Years of Potential Life Lost
Years of Potential Life Lost (YPLL) is another measure derived from mortality
data. It provides an estimate of the years lost due to premature death based
on a predetermined life expectancy. Deaths at a younger age have a greater impact
on this measure because more potential years of life are lost. As opposed to
age-specific or cause-specific mortality, this provides a measure of “prematurity”
of death.
(6) Impact of diseases on the change in life expectancy and Years of Life
Expectancy Lost
Based on a methodology developed by Arriaga using life expectancy at birth
and mortality data, it is also possible to measure the contribution of each
cause of death to the changes in life expectancy.(4) This recent innovation
in mortality analysis will be presented, along with the concept of Years of
Life Expectancy Lost (YLEL). YLEL, a different indicator than YPLL, uses life
expectancy to calculate the years lost in a population due to the mortality
experienced for such population.
(7) Disease-Adjusted Life Years (DALYs), Quality-Adjusted Life Years (QALYs)
The number of deaths alone can constitute indicators or they can be used
in conjunction with other health measures to construct new health indices. For
example, they may be combined with variables indicating the impact of premature
death and disability into disability-adjusted life years(5) (DALYs) or with
quality of life into quality-adjusted life years (QALYs).
The analysis of mortality data does not give a complete picture of the health
situation but it provides essential information about the health of a population.
Mortality measures have evolved and become more complex and encompass different
dimensions of health. In its most simple or most complex forms, they remain
important tools for situation analysis and decision-making in public health.
References:
(1) Rice University, Catalog of the Scientific Community [Internet site].
Available at: http://es.rice.edu/ES/humsoc/Galileo/Catalog/Files/graunt.html.
Accessed on 12/3/02.
(2) Pan American Health Organization, Area of Health Analysis and Information
Systems. Health Statistics from the Americas, 2003. [Internet site]. Available
soon at: http://www.paho.org.
(3) United Nations. The Millenium Assembly of the United Nations. New York,
NY:UN; 12 February 1999. Document A/RES/53/202.
(4) Arriaga EE. Los Años de Vida Perdidos: su utilización para
medir el nivel y cambio de la mortalidad. Notas de Población CELADE 1996;
24(63):7-38.
(5) Mathers CD, Vos T, Lopez AD, Salomon J, Ezzatti M (ed.) 2001. National Burden
of Disease Studies: A Practical Guide. Edition 2.0. Global Program on Evidence
for Health Policy. Geneva: World Health Organization.
Return to Index
Epidemiological Bulletin, Vol. 23 No. 4, Diciembre
2002